feat(analytics): add official provider ops metrics

Build on existing pricing funnel events with official provider, TTS usage, second-turn activation, and Flux balance bucket metrics.

apps/server/docs/ai-context/product-analytics-dashboard-setup.md

@@ -38,6 +38,9 @@ Live dashboard created on 2026-06-30:
 - Current cards:
   - Text card: `AIRI product analytics runbook`
   - Funnel: `Chat activation funnel`
+  - Trend: `Official provider usage`
+  - Funnel: `Official TTS activation`
+  - Trend: `Paywall exposure`
   - Trend: `Provider config failures`
   - Trend: `TTS voice selection and preview`
   - Trend: `Top selected TTS voices`
@@ -52,6 +55,7 @@ Steps:
 
 1. `chat_activation_started`
 2. `chat_activation_succeeded`
+3. `second_turn_started`
 
 Breakdowns:
 
@@ -67,6 +71,26 @@ Watch for:
 
 - Official provider conversion lower than custom provider conversion.
 - Large drop after `chat_activation_started`.
+- First-turn success but weak `second_turn_started` conversion.
+
+### Insight 1b: Official Provider Selection
+
+Type: Trends
+
+Events:
+
+- `official_provider_selected`
+
+Breakdowns:
+
+- `provider_id`
+- `source`
+- `auto_selected`
+
+Watch for:
+
+- Official provider auto-selection is present but users do not reach `second_turn_started`.
+- A single provider id dominates errors or activation drop-off.
 
 ### Insight 2: Chat Activation Failures
 
@@ -144,6 +168,10 @@ Events:
 - `voice_selected`
 - `voice_preview_played`
 - `voice_pack_bound`
+- `official_tts_exposed`
+- `official_tts_preview_started`
+- `official_tts_preview_succeeded`
+- `official_tts_auto_enabled`
 
 Breakdowns:
 
@@ -160,6 +188,11 @@ Use PostHog or SQL when grouping by:
 - `voice_id`
 - `voice_pack_id`
 
+Watch for:
+
+- Users see official TTS but do not preview it.
+- Official TTS preview succeeds but chat auto TTS is not triggered later.
+
 ### Insight 6: Voice Input Friction
 
 Type: Funnel
@@ -216,6 +249,8 @@ The `Product Analytics` row includes:
 - `Product Failure %`
 - `TTS Success %`
 - `TTS Failed / Blocked (range)`
+- `TTS Blocked by Reason`
+- `TTS Blocked by Flux Bucket`
 - `Top Product Actions (range)`
 - `Product Event Rate`
 - `TTS Event Rate by Source`
@@ -230,6 +265,8 @@ Live import status:
   - `Product Failure %`
   - `TTS Success %`
   - `TTS Failed / Blocked (range)`
+  - `TTS Blocked by Reason`
+  - `TTS Blocked by Flux Bucket`
   - `Top Product Actions (range)`
   - `Product Event Rate`
   - `TTS Event Rate by Source`
@@ -242,6 +279,7 @@ Permission notes from the import retry:
 - API confirmation returned `403 Access denied`: `You'll need additional permissions to perform this action. Permissions needed: any of dashboards:create, dashboards:write`.
 - The logged-in Grafana user `1260907335@qq.com` has org role `Viewer`; API metadata for `/d/rbr55dn/airi-server-overview` reports `canSave=false`, `canEdit=false`, `canAdmin=false`.
 - After permissions were updated, the generated dashboard was imported from Microsoft Edge. Grafana assigned the imported dashboard UID `ad8qbp5` instead of overwriting the earlier `rbr55dn` dashboard, so the imported dashboard was renamed to `AIRI Server Overview - Product Analytics` to avoid ambiguity.
+- On 2026-07-01, the live `ad8qbp5` dashboard was updated to include `TTS Blocked by Reason` and `TTS Blocked by Flux Bucket`. The live dashboard uses panel id `105` for the Flux bucket panel because id `103` was already occupied by the imported `User Engagement` row.
 
 Regenerate after dashboard changes:
 
@@ -298,8 +336,11 @@ clamp_min(sum(increase(airi_product_events_total{feature!="", action!=""}[15m]))
 
 ## Verification Checklist
 
-- PostHog can show `chat_activation_started -> chat_activation_succeeded` by `provider_mode`.
+- PostHog can show `chat_activation_started -> chat_activation_succeeded -> second_turn_started` by `provider_mode`.
+- PostHog can show `official_provider_selected` by `provider_id`, `source`, and `auto_selected`.
 - PostHog can show `voice_selected` by `voice_type` and `tts_provider_id`.
+- PostHog can show official TTS exposure / preview / auto-enabled events.
+- PostHog can show `paywall_seen` by `flux_balance_bucket`.
 - PostHog can show `feedback_submitted` and `bug_report_submitted`.
-- Grafana dashboard JSON contains `TTS Success %`, `TTS Failed / Blocked (range)`, and `TTS Event Rate by Source`.
-- Grafana product analytics panels use only bounded labels: `feature`, `action`, `status`, `source`.
+- Grafana dashboard JSON contains `TTS Success %`, `TTS Failed / Blocked (range)`, `TTS Blocked by Reason`, `TTS Blocked by Flux Bucket`, and `TTS Event Rate by Source`.
+- Grafana product analytics panels use only bounded labels: `feature`, `action`, `status`, `source`, `reason`, `flux_balance_bucket`.

apps/server/docs/ai-context/product-analytics-instrumentation.md

@@ -147,6 +147,8 @@ Grafana 当前不能回答：
 | `chat_activation_started` | frontend | PostHog | 用户进入首次聊天路径或点击发送第一条消息前 |
 | `chat_activation_succeeded` | frontend | PostHog | 首次消息完成并看到 assistant response |
 | `chat_activation_failed` | frontend | PostHog | 首次消息未完成，包含配置、网络、鉴权、余额、模型等失败 |
+| `official_provider_selected` | frontend | PostHog | 官方 Provider 被默认落地或在设置页被手动选择，记录 provider id 与是否自动选择 |
+| `second_turn_started` | frontend | PostHog | 同一会话开始第二轮对话 |
 
 字段：
 
@@ -159,6 +161,8 @@ Grafana 当前不能回答：
 | `time_to_first_message_ms` | success only | 从 app start 或 onboarding complete 到首次成功 |
 | `error_code` | failed only | 稳定错误码 |
 | `failure_stage` | failed only | `provider_config` / `model_list` / `message_send` / `llm_response` / `tts` |
+| `turn_index` | second turn only | 固定为 `2`，用于首轮成功后的二轮启动 |
+| `auto_selected` | official provider only | 官方默认 Provider 自动落地时为 `true` |
 
 推荐看板：
 
@@ -216,6 +220,9 @@ Grafana 当前不能回答：
 | `voice_selected` | frontend | PostHog | 用户选择音色，包括官方默认落地时的 baseline |
 | `voice_preview_played` | frontend | PostHog | 用户试听音色 |
 | `voice_pack_bound` | frontend | PostHog | 用户把 Voice Pack 绑定到角色 |
+| `official_tts_exposed` | frontend | PostHog | 官方 TTS 设置入口或激活入口被展示 |
+| `official_tts_preview_started` / `official_tts_preview_succeeded` | frontend | PostHog | 官方 TTS 试听开始 / 成功 |
+| `official_tts_auto_enabled` | frontend | PostHog | 聊天自动语音实际触发官方 TTS |
 | `speech_requested` | server | Postgres/Grafana | REST / WS TTS 请求开始；沿用现有 `feature = tts` action |
 | `speech_succeeded` | server | Postgres/Grafana | REST / WS TTS 交付成功；沿用现有 `feature = tts` action |
 | `speech_failed` | server | Postgres/Grafana | TTS 上游、配置、路由失败；沿用现有 `feature = tts` action |
@@ -456,6 +463,7 @@ Grafana 当前不能回答：
 | `tts_stop_clicked` | 保留用户停止行为 | 不代表音色选择；音色选择用 `voice_selected` |
 | `speech_requested` / `speech_succeeded` / `speech_failed` / `speech_blocked` | 服务端 TTS truth | 继续沿用 `feature = tts`，只补 metadata 字段 |
 | `pricing_page_viewed` / `plan_selected` / `checkout_started` | 保留付费漏斗前段 | 真正 payment completed 仍以 Stripe / Postgres 为准 |
+| `paywall_seen` | 付费漏斗入口 | 用 `flux_balance_bucket` 分层，不上报精确余额 |
 
 ## 数据卫生
 
@@ -859,15 +867,15 @@ Prometheus label 不放 `user_id`、`session_id`、`voice_pack_id`、自定义
 
 已接入代码：
 
-- Chat activation：`chat_activation_started`、`chat_activation_succeeded`、`chat_activation_failed`。
+- Chat activation：`chat_activation_started`、`chat_activation_succeeded`、`chat_activation_failed`、`second_turn_started`；官方 Provider 选择事件为 `official_provider_selected`，实际聊天使用口径看 activation events 的 `provider_mode = official`。
 - Model list：`model_list_loaded`、`model_list_failed`。
 - Provider config：`provider_config_started`、`provider_config_succeeded`、`provider_config_failed`。
-- TTS voice：`tts_provider_selected`、`voice_selected`、`voice_preview_played`、`voice_pack_bound`。
-- TTS 服务端 metadata：REST / WS TTS `product_events.metadata` 已补 `voice_id`、`voice_type`、`voice_pack_id`。
+- TTS voice：`tts_provider_selected`、`voice_selected`、`voice_preview_played`、`voice_pack_bound`、`official_tts_exposed`、`official_tts_preview_started`、`official_tts_preview_succeeded`、`official_tts_auto_enabled`。
+- TTS 服务端 metadata：REST / WS TTS `product_events.metadata` 已补 `voice_id`、`voice_type`、`voice_pack_id`、`block_reason`、`failure_reason`、`flux_balance_bucket`。
 - Voice input：`voice_input_started`、`microphone_permission_requested`、`microphone_permission_denied`、`audio_device_unavailable`、`voice_input_cancelled`。
 - STT：保留 `stt_started`、`stt_succeeded`、`stt_failed`，并将失败码收敛到稳定枚举，避免上报 raw error。
 - Feedback：`feedback_submitted` / `bug_report_submitted` 的低基数字段与 analytics API 已定义；产品内反馈提交入口与服务端收件流程拆到单独 PR。
-- Grafana Dashboard：`Product Analytics` 行已补 TTS success、TTS failed / blocked、TTS event rate by source 面板，并保留 voice drilldown 在 Postgres metadata / PostHog，不进入 Prometheus labels。
+- Grafana Dashboard：`Product Analytics` 行已补 TTS success、TTS failed / blocked、TTS event rate by source、TTS blocked by reason、TTS blocked by Flux bucket 面板，并保留 voice drilldown 在 Postgres metadata / PostHog，不进入 Prometheus labels。
 - Dashboard setup 文档：`product-analytics-dashboard-setup.md` 已补 PostHog insights、Grafana panels、PostHog / Grafana alert 配置建议。
 - 上线冒烟文档：`verifications/product-analytics-smoke.md` 已补 PostHog、Postgres、Grafana 三层验证步骤。
 

apps/server/docs/ai-context/verifications/product-analytics-smoke.md

@@ -1,8 +1,8 @@
 # Verification: Product Analytics Smoke Test
 
-Status: **code-level instrumentation verified; live PostHog dashboard created; Grafana dashboard imported; alert setup pending**
+Status: **code-level instrumentation verified; live PostHog dashboard updated; Grafana dashboard updated; alert setup pending**
 Owner: Community / Product Analytics
-Last updated: 2026-06-30
+Last updated: 2026-07-01
 Related:
 - [`product-analytics-instrumentation.md`](../product-analytics-instrumentation.md)
 - [`product-analytics-dashboard-setup.md`](../product-analytics-dashboard-setup.md)
@@ -12,7 +12,7 @@ Related:
 
 - **场景**：验证新增埋点能回答“用户是否能正常开始聊天”“Provider 配置卡在哪里”“哪个 TTS 音色被选择 / 实际播放”“语音输入卡在哪里”“用户是否提交反馈”。
 - **预期**：PostHog 能看到前端 journey events；Postgres `product_events` 能看到服务端 TTS metadata；Grafana 能看到低基数 server-side product health。
-- **当前状态**：代码与 dashboard JSON 已验证；线上 PostHog dashboard 已创建；线上 Grafana `AIRI Server Overview - Product Analytics` (`ad8qbp5`) 已导入完整 Product Analytics row；alert 仍需人工配置。
+- **当前状态**：代码与 dashboard JSON 已验证；线上 PostHog dashboard 已补官方 Provider / 官方 TTS / paywall 卡片；线上 Grafana `AIRI Server Overview - Product Analytics` (`ad8qbp5`) 已补 TTS blocked reason / Flux bucket 面板；alert 仍需人工配置。
 
 ## 已经由代码验证
 
@@ -35,12 +35,14 @@ Action:
 1. Use a fresh or test account.
 2. Start with an official provider.
 3. Send the first chat message and wait for the assistant response.
+4. Send a second message in the same session.
 
 Expected PostHog events:
 
 ```text
 chat_activation_started
 chat_activation_succeeded
+second_turn_started
 ```
 
 Required properties:
@@ -50,14 +52,42 @@ provider_mode = official
 provider_id = <official provider id>
 model_id = <selected model id>
 surface = web | mobile | electron
+turn_index = 2
 ```
 
 Fail if:
 
 - `chat_activation_started` appears but `chat_activation_succeeded` never appears for a successful chat.
+- The second message is sent but `second_turn_started` does not appear.
 - `provider_mode` is missing or always `unknown`.
 - `surface` is missing.
 
+### 1b. PostHog: official provider selection
+
+Action:
+
+1. Sign in with an account that has no active chat provider yet, or switch the chat provider to the official provider in settings.
+
+Expected PostHog events:
+
+```text
+official_provider_selected
+```
+
+Required properties:
+
+```text
+provider_mode = official
+provider_id = <official provider id>
+source = default_auto | settings
+auto_selected = true | false
+```
+
+Fail if:
+
+- Default official provider bootstrap reports `auto_selected = false`.
+- Manual settings selection reports `auto_selected = true`.
+
 ### 2. PostHog: provider config failure
 
 Action:
@@ -99,6 +129,9 @@ Expected PostHog events:
 
 ```text
 tts_provider_selected
+official_tts_exposed
+official_tts_preview_started
+official_tts_preview_succeeded
 voice_selected
 voice_preview_played
 ```
@@ -116,8 +149,35 @@ source = settings | manual_preview
 Fail if:
 
 - `voice_selected` is missing, because this blocks “哪个 TTS 音色比较多”的核心问题。
+- Official TTS preview succeeds in the UI but `official_tts_preview_succeeded` is missing.
 - Official default voice is indistinguishable from custom configured voice.
 
+### 3b. PostHog: official TTS auto playback
+
+Action:
+
+1. Enable chat auto TTS with an official TTS provider.
+2. Send a chat message and wait for an assistant response that triggers speech playback.
+
+Expected PostHog events:
+
+```text
+official_tts_auto_enabled
+```
+
+Required properties:
+
+```text
+tts_provider_id = <official provider id>
+tts_model_id = <model id>
+source = chat_auto_tts
+enabled = true
+```
+
+Fail if:
+
+- Chat auto TTS plays through the official provider but `official_tts_auto_enabled` is missing.
+
 ### 4. PostHog: voice input friction
 
 Action:
@@ -143,6 +203,32 @@ Fail if:
 - Permission denied or device unavailable is only visible as a generic `stt_failed`.
 - `error_code` contains raw browser error text.
 
+### 4b. PostHog: paywall exposure
+
+Action:
+
+1. Open the Flux / plan purchase entry.
+2. Use an account with a known low or zero Flux balance if possible.
+
+Expected PostHog events:
+
+```text
+paywall_seen
+```
+
+Required properties:
+
+```text
+surface = settings_flux
+reason = manual_topup
+flux_balance_bucket = zero | 1_100 | 101_1000 | 1001_10000 | 10000_plus | unknown
+```
+
+Fail if:
+
+- The purchase entry is visible but `paywall_seen` is missing.
+- A precise balance is sent instead of the bounded `flux_balance_bucket`.
+
 ### 5. Postgres: server-side TTS metadata
 
 Action:
@@ -176,6 +262,8 @@ Expected:
 - `speech_requested` and `speech_succeeded` rows exist for successful TTS.
 - `voice_id` is present when the request provided a selected voice.
 - `voice_type` distinguishes official default / selected / custom / voice pack where available.
+- Blocked rows include bounded `block_reason` and `flux_balance_bucket`.
+- Failed rows include bounded `failure_reason`.
 
 Fail if:
 
@@ -197,6 +285,8 @@ Product Events (range)
 Product Failure %
 TTS Success %
 TTS Failed / Blocked (range)
+TTS Blocked by Reason
+TTS Blocked by Flux Bucket
 Top Product Actions (range)
 Product Event Rate
 TTS Event Rate by Source
@@ -211,7 +301,7 @@ sum(increase(airi_product_events_total{feature="tts"}[1h]))
 Expected:
 
 - Query returns a non-zero value after TTS smoke actions.
-- Legends only use bounded labels: `feature`, `action`, `status`, `source`.
+- Legends only use bounded labels: `feature`, `action`, `status`, `source`, `reason`, `flux_balance_bucket`.
 
 Fail if:
 
@@ -262,14 +352,16 @@ ORDER BY event_count DESC;
 | Item | Pass condition |
 |---|---|
 | Activation | PostHog funnel shows `chat_activation_started -> chat_activation_succeeded` by `provider_mode` |
+| Retention proxy | PostHog shows `second_turn_started` for the second message in a successful session |
+| Official provider | PostHog shows `official_provider_selected` by `provider_id` and `source` |
 | Provider config | Failed custom config emits `provider_config_failed` with bounded `error_code` |
-| TTS voice | PostHog can rank `voice_selected` by `voice_id`; Postgres can rank actual `speech_succeeded` by metadata voice |
+| TTS voice | PostHog can rank `voice_selected` by `voice_id`; official TTS exposure / preview / auto playback events appear; Postgres can rank actual `speech_succeeded` by metadata voice |
 | Voice input | Permission / device / cancel paths are distinguishable |
 | Feedback | Feedback event API exists with bounded fields; product feedback UI/server submission is split into a separate PR |
-| Grafana | Product Analytics row renders and uses only bounded Prometheus labels |
+| Grafana | Product Analytics row renders TTS reason / Flux bucket panels and uses only bounded Prometheus labels |
 
 ## Known Pending Work
 
-- PostHog dashboard and alerts still need to be created inside the PostHog account.
-- Updated Grafana JSON still needs to be imported or deployed to the production Grafana workspace.
+- PostHog dashboard cards are created, but the official provider / official TTS / paywall cards need deployed traffic before they show meaningful data.
+- Updated Grafana panels are deployed to the production Grafana workspace, but alert rules still need to be configured.
 - Discord / QQ ingestion and daily / weekly automation scripts are intentionally excluded from this pass.

apps/server/otel/grafana/dashboards/airi-server-overview-cloud.json

@@ -1562,6 +1562,172 @@
         }
       }
     },
+    "panel-102": {
+      "kind": "Panel",
+      "spec": {
+        "data": {
+          "kind": "QueryGroup",
+          "spec": {
+            "queries": [
+              {
+                "kind": "PanelQuery",
+                "spec": {
+                  "hidden": false,
+                  "query": {
+                    "datasource": {
+                      "name": "grafanacloud-projairi-prom"
+                    },
+                    "group": "prometheus",
+                    "kind": "DataQuery",
+                    "spec": {
+                      "expr": "topk(12, sum by (reason, source) (increase(airi_product_events_total{service_name=~\"$service\", deployment_environment=~\"$env\", feature!=\"\", action!=\"\", feature=\"tts\", action=\"speech_blocked\", status=\"blocked\"}[$__range])))",
+                      "legendFormat": "{{reason}} · {{source}}",
+                      "instant": true,
+                      "range": false
+                    },
+                    "version": "v0"
+                  },
+                  "refId": "A"
+                }
+              }
+            ],
+            "queryOptions": {},
+            "transformations": []
+          }
+        },
+        "description": "Blocked TTS events over the dashboard range, grouped by bounded product reason and source. Today this mostly shows insufficient balance; new policy/provider/preflight buckets can be added without exposing user, voice, or model labels.",
+        "id": 102,
+        "links": [],
+        "title": "TTS Blocked by Reason",
+        "vizConfig": {
+          "group": "bargauge",
+          "kind": "VizConfig",
+          "spec": {
+            "fieldConfig": {
+              "defaults": {
+                "color": {
+                  "mode": "thresholds"
+                },
+                "thresholds": {
+                  "mode": "absolute",
+                  "steps": [
+                    {
+                      "color": "green",
+                      "value": 0
+                    }
+                  ]
+                },
+                "unit": "short",
+                "noValue": "0"
+              },
+              "overrides": []
+            },
+            "options": {
+              "displayMode": "gradient",
+              "maxVizHeight": 300,
+              "minVizHeight": 12,
+              "minVizWidth": 8,
+              "namePlacement": "auto",
+              "orientation": "horizontal",
+              "reduceOptions": {
+                "calcs": [
+                  "lastNotNull"
+                ],
+                "fields": "",
+                "values": false
+              },
+              "showUnfilled": true,
+              "sizing": "auto",
+              "valueMode": "color"
+            }
+          },
+          "version": "13.0.0-23630096546"
+        }
+      }
+    },
+    "panel-103": {
+      "kind": "Panel",
+      "spec": {
+        "data": {
+          "kind": "QueryGroup",
+          "spec": {
+            "queries": [
+              {
+                "kind": "PanelQuery",
+                "spec": {
+                  "hidden": false,
+                  "query": {
+                    "datasource": {
+                      "name": "grafanacloud-projairi-prom"
+                    },
+                    "group": "prometheus",
+                    "kind": "DataQuery",
+                    "spec": {
+                      "expr": "topk(8, sum by (flux_balance_bucket, source) (increase(airi_product_events_total{service_name=~\"$service\", deployment_environment=~\"$env\", feature!=\"\", action!=\"\", feature=\"tts\", action=\"speech_blocked\", status=\"blocked\", flux_balance_bucket!=\"\"}[$__range])))",
+                      "legendFormat": "{{flux_balance_bucket}} · {{source}}",
+                      "instant": true,
+                      "range": false
+                    },
+                    "version": "v0"
+                  },
+                  "refId": "A"
+                }
+              }
+            ],
+            "queryOptions": {},
+            "transformations": []
+          }
+        },
+        "description": "Blocked TTS events over the dashboard range, grouped by coarse Flux balance bucket. This helps separate truly empty accounts from low-balance accounts without exposing exact user balances.",
+        "id": 103,
+        "links": [],
+        "title": "TTS Blocked by Flux Bucket",
+        "vizConfig": {
+          "group": "bargauge",
+          "kind": "VizConfig",
+          "spec": {
+            "fieldConfig": {
+              "defaults": {
+                "color": {
+                  "mode": "thresholds"
+                },
+                "thresholds": {
+                  "mode": "absolute",
+                  "steps": [
+                    {
+                      "color": "green",
+                      "value": 0
+                    }
+                  ]
+                },
+                "unit": "short",
+                "noValue": "0"
+              },
+              "overrides": []
+            },
+            "options": {
+              "displayMode": "gradient",
+              "maxVizHeight": 300,
+              "minVizHeight": 12,
+              "minVizWidth": 8,
+              "namePlacement": "auto",
+              "orientation": "horizontal",
+              "reduceOptions": {
+                "calcs": [
+                  "lastNotNull"
+                ],
+                "fields": "",
+                "values": false
+              },
+              "showUnfilled": true,
+              "sizing": "auto",
+              "valueMode": "color"
+            }
+          },
+          "version": "13.0.0-23630096546"
+        }
+      }
+    },
     "panel-16": {
       "kind": "Panel",
       "spec": {
@@ -4564,6 +4730,32 @@
                       "x": 12,
                       "y": 9
                     }
+                  },
+                  {
+                    "kind": "GridLayoutItem",
+                    "spec": {
+                      "element": {
+                        "kind": "ElementReference",
+                        "name": "panel-102"
+                      },
+                      "height": 5,
+                      "width": 12,
+                      "x": 0,
+                      "y": 13
+                    }
+                  },
+                  {
+                    "kind": "GridLayoutItem",
+                    "spec": {
+                      "element": {
+                        "kind": "ElementReference",
+                        "name": "panel-103"
+                      },
+                      "height": 5,
+                      "width": 12,
+                      "x": 12,
+                      "y": 13
+                    }
                   }
                 ]
               }

apps/server/otel/grafana/dashboards/build.test.ts

@@ -64,6 +64,8 @@ describe('grafana dashboard builder', () => {
     expect(panelTitle('panel-99')).toBe('TTS Success %')
     expect(panelTitle('panel-100')).toBe('TTS Failed / Blocked (range)')
     expect(panelTitle('panel-101')).toBe('TTS Event Rate by Source')
+    expect(panelTitle('panel-102')).toBe('TTS Blocked by Reason')
+    expect(panelTitle('panel-103')).toBe('TTS Blocked by Flux Bucket')
   })
 
   /**
@@ -80,6 +82,8 @@ describe('grafana dashboard builder', () => {
       dashboard.elements['panel-99'],
       dashboard.elements['panel-100'],
       dashboard.elements['panel-101'],
+      dashboard.elements['panel-102'],
+      dashboard.elements['panel-103'],
     ]).join('\n')
 
     expect(productPanelExpressions).not.toContain('voice_id')

apps/server/otel/grafana/dashboards/build.ts

@@ -613,6 +613,34 @@ elements['panel-101'] = timeseriesPanel(
   { unit: 'eps', fillOpacity: 15 },
 )
 
+elements['panel-102'] = barGaugePanel(
+  102,
+  'TTS Blocked by Reason',
+  'Blocked TTS events over the dashboard range, grouped by bounded product reason and source. Today this mostly shows insufficient balance; new policy/provider/preflight buckets can be added without exposing user, voice, or model labels.',
+  [query(
+    `topk(12, sum by (reason, source) (increase(airi_product_events_total{${PRODUCT_EVENT_FILTER}, feature="tts", action="speech_blocked", status="blocked"}[$__range])))`,
+    '{{reason}} · {{source}}',
+    'A',
+    PROM,
+    { instant: true },
+  )],
+  { unit: 'short', noValue: '0' },
+)
+
+elements['panel-103'] = barGaugePanel(
+  103,
+  'TTS Blocked by Flux Bucket',
+  'Blocked TTS events over the dashboard range, grouped by coarse Flux balance bucket. This helps separate truly empty accounts from low-balance accounts without exposing exact user balances.',
+  [query(
+    `topk(8, sum by (flux_balance_bucket, source) (increase(airi_product_events_total{${PRODUCT_EVENT_FILTER}, feature="tts", action="speech_blocked", status="blocked", flux_balance_bucket!=""}[$__range])))`,
+    '{{flux_balance_bucket}} · {{source}}',
+    'A',
+    PROM,
+    { instant: true },
+  )],
+  { unit: 'short', noValue: '0' },
+)
+
 // --- Row 2: HTTP — traffic ranking, error trend, latency trend -------------
 elements['panel-16'] = barGaugePanel(
   16,
@@ -955,6 +983,8 @@ const rows = [
     item('panel-97', 0, 5, 12, 8),
     item('panel-98', 12, 5, 12, 4),
     item('panel-101', 12, 9, 12, 4),
+    item('panel-102', 0, 13, 12, 5),
+    item('panel-103', 12, 13, 12, 5),
   ]),
   // Row 3: HTTP — full-width error breakdown on top, then traffic ranking +
   // latency trend side by side.

apps/server/src/routes/audio-speech-ws/route.test.ts

@@ -322,7 +322,9 @@ describe('audio-speech-ws route', () => {
       reason: 'insufficient_balance',
       metadata: expect.objectContaining({
         trigger: 'auto',
+        block_reason: 'insufficient_balance',
         balance_state: 'insufficient',
+        flux_balance_bucket: 'zero',
       }),
     }))
   })

apps/server/src/routes/audio-speech-ws/session.ts

@@ -11,6 +11,7 @@ import WebSocket from 'ws'
 import { useLogger } from '@guiiai/logg'
 import { context as otelContext, SpanStatusCode, trace } from '@opentelemetry/api'
 
+import { fluxBalanceBucket } from '../../services/domain/flux-balance'
 import { ApiError } from '../../utils/error'
 import { nanoid } from '../../utils/id'
 import {
@@ -94,6 +95,7 @@ export function createSessionState(
   let closed = false
   let billed = false
   let totalInputChars = 0
+  let preflightFluxBalance: number | undefined
   let modelLabel = STREAM_MODEL_LABEL_FALLBACK
   let voiceLabel: string | undefined
   /**
@@ -141,6 +143,7 @@ export function createSessionState(
     // afford the worst-case session.
     try {
       const flux = await opts.fluxService.getFlux(userId)
+      preflightFluxBalance = flux.flux
       await opts.ttsMeter.assertCanAfford(userId, STREAMING_PREFLIGHT_CHARS_ESTIMATE, flux.flux)
     }
     catch (err) {
@@ -508,7 +511,9 @@ export function createSessionState(
       model: modelLabel,
       reason: 'insufficient_balance',
       metadata: {
+        block_reason: 'insufficient_balance',
         balance_state: 'insufficient',
+        flux_balance_bucket: fluxBalanceBucket(preflightFluxBalance),
         billing_units: STREAMING_PREFLIGHT_CHARS_ESTIMATE,
         close_code: code,
         duration_ms: Date.now() - startedAt,

apps/server/src/routes/openai/v1/route.test.ts

@@ -975,6 +975,7 @@ describe('v1CompletionsRoutes', () => {
           action: 'speech_failed',
           reason: 'TOO_MANY_REQUESTS',
           metadata: expect.objectContaining({
+            failure_reason: 'TOO_MANY_REQUESTS',
             http_status: 429,
           }),
         }),
@@ -1012,7 +1013,9 @@ describe('v1CompletionsRoutes', () => {
         reason: 'insufficient_balance',
         metadata: expect.objectContaining({
           trigger: 'manual',
+          block_reason: 'insufficient_balance',
           balance_state: 'insufficient',
+          flux_balance_bucket: 'zero',
         }),
       }))
     })
@@ -1058,7 +1061,9 @@ describe('v1CompletionsRoutes', () => {
         reason: 'insufficient_balance',
         metadata: expect.objectContaining({
           trigger: 'auto',
+          block_reason: 'insufficient_balance',
           balance_state: 'insufficient',
+          flux_balance_bucket: 'zero',
         }),
       }))
     })

apps/server/src/services/domain/flux-balance.ts

@@ -0,0 +1,28 @@
+export type FluxBalanceBucket = 'zero' | '1_100' | '101_1000' | '1001_10000' | '10000_plus' | 'unknown'
+
+/**
+ * Normalizes exact Flux balance values into analytics-safe buckets.
+ *
+ * Before:
+ * - 0
+ * - 42
+ * - 1200
+ *
+ * After:
+ * - "zero"
+ * - "1_100"
+ * - "1001_10000"
+ */
+export function fluxBalanceBucket(balance: number | null | undefined): FluxBalanceBucket {
+  if (balance == null || Number.isNaN(balance))
+    return 'unknown'
+  if (balance <= 0)
+    return 'zero'
+  if (balance <= 100)
+    return '1_100'
+  if (balance <= 1000)
+    return '101_1000'
+  if (balance <= 10000)
+    return '1001_10000'
+  return '10000_plus'
+}

apps/server/src/services/domain/openai-speech/index.ts

@@ -18,6 +18,7 @@ import {
   AIRI_ATTR_GEN_AI_OPERATION_KIND,
   GEN_AI_ATTR_REQUEST_MODEL,
 } from '../../../utils/observability'
+import { fluxBalanceBucket } from '../flux-balance'
 
 const tracer = trace.getTracer('v1-completions')
 
@@ -153,7 +154,9 @@ export function createOpenAiSpeechService(deps: OpenAiSpeechServiceDeps) {
         metadata: {
           input_chars: inputText.length,
           billing_units: billingUnits,
+          block_reason: 'insufficient_balance',
           balance_state: 'insufficient',
+          flux_balance_bucket: fluxBalanceBucket(flux.flux),
           trigger: analytics.trigger,
           ...voiceMetadata,
         },
@@ -230,6 +233,7 @@ export function createOpenAiSpeechService(deps: OpenAiSpeechServiceDeps) {
         metadata: {
           http_status: failure.status,
           duration_ms: Date.now() - startedAt,
+          failure_reason: failure.reason,
           trigger: analytics.trigger,
           ...voiceMetadata,
         },
@@ -257,6 +261,7 @@ export function createOpenAiSpeechService(deps: OpenAiSpeechServiceDeps) {
         metadata: {
           http_status: response.status,
           duration_ms: durationMs,
+          failure_reason: 'upstream_error',
           trigger: analytics.trigger,
           ...voiceMetadata,
         },

apps/server/src/services/domain/product-events.test.ts

@@ -114,6 +114,7 @@ describe('productEventService', () => {
       metadata: {
         trigger: 'auto',
         balance_state: 'insufficient',
+        flux_balance_bucket: 'zero',
       },
     })
 
@@ -131,6 +132,8 @@ describe('productEventService', () => {
       action: 'speech_blocked',
       status: 'blocked',
       source: 'chat_auto_tts',
+      reason: 'insufficient_balance',
+      flux_balance_bucket: 'zero',
     })
   })
 })

apps/server/src/services/domain/product-events.ts

@@ -75,6 +75,27 @@ export interface ProductEventAggregateRow {
   distinctUsers: number
 }
 
+/**
+ * Builds bounded Prometheus labels from product event inputs.
+ */
+function metricLabels(input: ProductEventInput): Record<string, string> {
+  const attrs: Record<string, string> = {
+    feature: input.feature,
+    action: input.action,
+    status: input.status,
+  }
+  if (input.source)
+    attrs.source = input.source
+  if (input.reason)
+    attrs.reason = input.reason
+
+  const fluxBalanceBucket = input.metadata?.flux_balance_bucket
+  if (typeof fluxBalanceBucket === 'string')
+    attrs.flux_balance_bucket = fluxBalanceBucket
+
+  return attrs
+}
+
 /**
  * Creates AIRI's first-party product analytics event writer.
  *
@@ -109,14 +130,7 @@ export function createProductEventService(db: Database, metrics?: ProductMetrics
           createdAt: input.createdAt,
         })
 
-        const attrs: Record<string, string> = {
-          feature: input.feature,
-          action: input.action,
-          status: input.status,
-        }
-        if (input.source)
-          attrs.source = input.source
-        metrics?.events.add(1, attrs)
+        metrics?.events.add(1, metricLabels(input))
       }
       catch (err) {
         logger.withError(err).withFields({

packages/core-agent/src/runtime/chat-orchestrator-runtime.test.ts

@@ -332,6 +332,7 @@ describe('createChatOrchestratorRuntime', () => {
     expect(harness.telemetry.chatActivationStarted).toEqual([{
       model: 'gpt-test',
       provider: 'mock-provider',
+      sessionId: 'session-1',
       source: 'voice',
     }])
     expect(harness.telemetry.chatActivationSucceeded).toEqual([{
@@ -359,6 +360,7 @@ describe('createChatOrchestratorRuntime', () => {
     expect(harness.telemetry.chatActivationStarted).toEqual([{
       model: 'gpt-test',
       provider: 'mock-provider',
+      sessionId: 'session-1',
       source: 'text',
     }])
     expect(harness.telemetry.chatActivationSucceeded).toEqual([])

packages/core-agent/src/runtime/chat-orchestrator-runtime.ts

@@ -192,6 +192,7 @@ export interface ChatOrchestratorRuntimeDeps {
   onTrackFirstMessage?: () => void
   /** Called when a user starts a chat activation attempt. */
   onChatActivationStarted?: (event: {
+    sessionId: string
     source: 'text' | 'voice'
     model: string
     provider: string
@@ -247,6 +248,10 @@ export interface ChatOrchestratorRuntimeDeps {
     sessionId: string
     message: Extract<ChatHistoryItem, { role: 'user' }> & { id: string }
     messageText: string
+    source: 'text' | 'voice'
+    model: string
+    provider: string
+    turnIndex: number
   }) => void
   /** Called after the assistant message has been finalized into session history. */
   onAssistantMessageAppended?: (event: {
@@ -426,6 +431,7 @@ export function createChatOrchestratorRuntime(deps: ChatOrchestratorRuntimeDeps)
     const activeProvider = deps.getActiveProvider?.() ?? ''
     deps.onTrackFirstMessage?.()
     deps.onChatActivationStarted?.({
+      sessionId,
       source: sendSource,
       model: options.model,
       provider: activeProvider,
@@ -475,6 +481,7 @@ export function createChatOrchestratorRuntime(deps: ChatOrchestratorRuntimeDeps)
         id: userMessageId,
       }
       deps.session.appendSessionMessage(sessionId, userMessage)
+      const userTurnIndex = deps.session.getSessionMessages(sessionId).filter(message => message.role === 'user').length
 
       // Cloud sync v1: only the raw text part round-trips; image attachments
       // and other non-text parts stay local.
@@ -482,6 +489,10 @@ export function createChatOrchestratorRuntime(deps: ChatOrchestratorRuntimeDeps)
         sessionId,
         message: userMessage,
         messageText: sendingMessage,
+        source: sendSource,
+        model: options.model,
+        provider: activeProvider,
+        turnIndex: userTurnIndex,
       })
 
       const sessionMessagesForSend = deps.session.getSessionMessages(sessionId)

packages/stage-pages/src/pages/settings/flux.vue

@@ -1,4 +1,6 @@
 <script setup lang="ts">
+import type { FluxBalanceBucket } from '@proj-airi/stage-ui/composables/use-analytics'
+
 import { isFluxPurchaseDisabled, isStageTamagotchi } from '@proj-airi/stage-shared'
 import { client } from '@proj-airi/stage-ui/composables/api'
 import { useAnalytics } from '@proj-airi/stage-ui/composables/use-analytics'
@@ -17,6 +19,7 @@ const authStore = useAuthStore()
 const { credits } = storeToRefs(authStore)
 const {
   trackCheckoutStarted,
+  trackPaywallSeen,
   trackPlanSelected,
   trackPricingViewed,
   trackQuotaLimitReached,
@@ -42,6 +45,7 @@ interface FluxPackage {
 
 const loadingPriceId = ref<string | null>(null)
 const message = ref<{ type: 'success' | 'error', text: string } | null>(null)
+const checkoutReturnMessageActive = ref(false)
 const packages = ref<FluxPackage[]>([])
 const selectedCurrency = ref<string>('usd')
 
@@ -71,6 +75,23 @@ function formatNumber(num: number): string {
   return new Intl.NumberFormat().format(num)
 }
 
+/**
+ * Buckets Flux balances so monetization analytics never expose exact balances.
+ */
+function fluxBalanceBucket(balance: number | undefined): FluxBalanceBucket {
+  if (balance == null || Number.isNaN(balance))
+    return 'unknown'
+  if (balance <= 0)
+    return 'zero'
+  if (balance <= 100)
+    return '1_100'
+  if (balance <= 1000)
+    return '101_1000'
+  if (balance <= 10000)
+    return '1001_10000'
+  return '10000_plus'
+}
+
 /** Display amount with sign: debit is negative, credit/initial are positive */
 function displayAmount(record: AuditRecord): string {
   const signed = record.type === 'debit' ? -record.amount : record.amount
@@ -233,18 +254,44 @@ async function fetchPackages() {
     }
   }
   catch {
-    message.value = { type: 'error', text: t('settings.pages.flux.packagesError') }
+    if (!checkoutReturnMessageActive.value)
+      message.value = { type: 'error', text: t('settings.pages.flux.packagesError') }
   }
 }
 
+/**
+ * Shows a Stripe return banner that background package refreshes must not replace.
+ */
+function showCheckoutReturnMessage(type: 'success' | 'error', text: string) {
+  checkoutReturnMessageActive.value = true
+  message.value = { type, text }
+}
+
 onMounted(async () => {
-  await Promise.allSettled([authStore.updateCredits(), fetchStats(), fetchAuditHistory(), ...(fluxPurchaseDisabled ? [] : [fetchPackages()])])
+  const creditsRefresh = authStore.updateCredits()
+  void Promise.allSettled([fetchStats(), fetchAuditHistory(), ...(fluxPurchaseDisabled ? [] : [fetchPackages()])])
+
+  if (route.query.success === 'true') {
+    showCheckoutReturnMessage('success', t('settings.pages.flux.checkout.success'))
+    router.replace({ query: {} })
+  }
+  else if (route.query.canceled === 'true') {
+    showCheckoutReturnMessage('error', t('settings.pages.flux.checkout.canceled'))
+    router.replace({ query: {} })
+  }
+
+  await creditsRefresh.catch(() => undefined)
 
   // PostHog funnel step 1: pricing surface view. Today this is an in-app
   // settings page (already-authenticated users); when we add a public
   // pricing landing page the surface label changes but the event stays the
   // same, so the funnel definition in PostHog doesn't need re-wiring.
   if (!fluxPurchaseDisabled) {
+    trackPaywallSeen({
+      surface: 'settings_flux',
+      reason: 'manual_topup',
+      flux_balance_bucket: fluxBalanceBucket(credits.value),
+    })
     trackPricingViewed('settings_flux', 'one_time')
     if (credits.value <= 0) {
       trackQuotaLimitReached({
@@ -255,19 +302,11 @@ onMounted(async () => {
       })
     }
   }
-
-  if (route.query.success === 'true') {
-    message.value = { type: 'success', text: t('settings.pages.flux.checkout.success') }
-    router.replace({ query: {} })
-  }
-  else if (route.query.canceled === 'true') {
-    message.value = { type: 'error', text: t('settings.pages.flux.checkout.canceled') }
-    router.replace({ query: {} })
-  }
 })
 
 async function handleBuy(stripePriceId: string) {
   loadingPriceId.value = stripePriceId
+  checkoutReturnMessageActive.value = false
   message.value = null
   // PostHog funnel step 2: user picked a plan. price_minor_unit lives on
   // the Stripe webhook (server-side `payment_completed`); we deliberately

packages/stage-pages/src/pages/settings/modules/consciousness.vue

@@ -25,7 +25,23 @@ const {
 } = storeToRefs(consciousnessStore)
 
 const { t } = useI18n()
-const { trackProviderClick } = useAnalytics()
+const { trackOfficialProviderSelected, trackProviderClick } = useAnalytics()
+
+/**
+ * Tracks explicit official chat-provider selection from settings.
+ */
+function trackOfficialProviderSelection(providerId: string, modelId: string) {
+  if (!providerId.startsWith('official-provider'))
+    return
+
+  trackOfficialProviderSelected({
+    provider_id: providerId,
+    provider_mode: 'official',
+    source: 'settings',
+    auto_selected: false,
+    model_id: modelId || 'unknown',
+  })
+}
 
 watch(activeProvider, async (provider, oldProvider) => {
   if (!provider)
@@ -34,6 +50,7 @@ watch(activeProvider, async (provider, oldProvider) => {
   // Reset model when switching providers (but not on initial load)
   if (oldProvider !== undefined && oldProvider !== provider) {
     activeModel.value = ''
+    trackOfficialProviderSelection(provider, activeModel.value)
   }
 
   await consciousnessStore.loadModelsForProvider(provider)

packages/stage-pages/src/pages/settings/modules/speech.vue

@@ -53,6 +53,9 @@ const {
 
 const {
   trackProviderClick,
+  trackOfficialTtsExposed,
+  trackOfficialTtsPreviewStarted,
+  trackOfficialTtsPreviewSucceeded,
   trackTtsProviderSelected,
   trackVoicePreviewPlayed,
   trackVoiceSelected,
@@ -66,6 +69,7 @@ const isGenerating = ref(false)
 const audioUrl = ref('')
 const audioPlayer = ref<HTMLAudioElement | null>(null)
 const errorMessage = ref('')
+let lastOfficialTtsExposureKey = ''
 
 const STREAMING_MODEL_OPTION_PREFIX = 'streaming:'
 
@@ -194,6 +198,13 @@ function currentTtsModelId() {
   return activeSpeechModel.value || 'unknown'
 }
 
+/**
+ * Checks whether the selected provider is one of AIRI's official TTS providers.
+ */
+function isOfficialTtsProvider(providerId: string) {
+  return providerId === OFFICIAL_SPEECH_PROVIDER_ID || providerId === OFFICIAL_SPEECH_STREAMING_PROVIDER_ID
+}
+
 /**
  * Classifies the selected voice without sending free-form provider config as a dimension.
  */
@@ -299,6 +310,25 @@ function selectSpeechModel(modelOptionId: string) {
   activeSpeechModel.value = nextModel
 }
 
+/**
+ * Tracks that the settings page has shown an official TTS route to the user.
+ */
+function trackOfficialTtsExposure(providerId = activeSpeechProvider.value, modelId = currentTtsModelId()) {
+  if (!providerId || !isOfficialTtsProvider(providerId))
+    return
+
+  const exposureKey = `${providerId}:${modelId}`
+  if (lastOfficialTtsExposureKey === exposureKey)
+    return
+
+  lastOfficialTtsExposureKey = exposureKey
+  trackOfficialTtsExposed({
+    tts_provider_id: providerId,
+    tts_model_id: modelId,
+    source: 'settings',
+  })
+}
+
 // Sync OpenAI Compatible model and voice from provider config
 function syncOpenAICompatibleSettings() {
   if (activeSpeechProvider.value !== 'openai-compatible-audio-speech')
@@ -331,6 +361,7 @@ onMounted(async () => {
   speechStore.ensureActiveSpeechModel()
   await speechStore.loadVoicesForProvider(activeSpeechProvider.value, activeSpeechModel.value || undefined)
   syncOpenAICompatibleSettings()
+  trackOfficialTtsExposure()
 })
 
 watch(activeSpeechProvider, async (newProvider, oldProvider) => {
@@ -355,6 +386,7 @@ watch(activeSpeechProvider, async (newProvider, oldProvider) => {
   // explicit ?model=). No-op for other providers / when a model is selected.
   speechStore.ensureActiveSpeechModel()
   await speechStore.loadVoicesForProvider(newProvider, activeSpeechModel.value || undefined)
+  trackOfficialTtsExposure(newProvider, currentTtsModelId())
 
   syncOpenAICompatibleSettings()
 })
@@ -367,6 +399,7 @@ watch(activeSpeechModel, async (model) => {
   activeSpeechVoice.value = undefined
 
   await speechStore.loadVoicesForProvider(activeSpeechProvider.value, model || undefined)
+  trackOfficialTtsExposure(activeSpeechProvider.value, currentTtsModelId())
 })
 
 watch([activeSpeechProvider, activeSpeechModel, activeSpeechVoiceId], ([provider, model, voiceId]) => {
@@ -424,6 +457,7 @@ async function generateTestSpeech() {
   const previewModel = model
   const previewProvider = activeSpeechProvider.value || 'unknown'
   const previewAnalytics = voiceAnalyticsPayload(previewVoice.id, previewProvider)
+  const previewStartedAt = performance.now()
 
   isGenerating.value = true
   errorMessage.value = ''
@@ -450,6 +484,15 @@ async function generateTestSpeech() {
           supportsSSML: speechStore.supportsSSML,
         })
 
+    if (isOfficialTtsProvider(previewProvider)) {
+      trackOfficialTtsPreviewStarted({
+        tts_provider_id: previewProvider,
+        tts_model_id: previewModel,
+        ...previewAnalytics,
+        source: 'manual_preview',
+      })
+    }
+
     const response = await generateSpeech({
       ...provider.speech(
         model,
@@ -461,6 +504,15 @@ async function generateTestSpeech() {
 
     // Convert the response to a blob and create an object URL
     audioUrl.value = URL.createObjectURL(new Blob([response]))
+    if (isOfficialTtsProvider(previewProvider)) {
+      trackOfficialTtsPreviewSucceeded({
+        tts_provider_id: previewProvider,
+        tts_model_id: previewModel,
+        ...previewAnalytics,
+        source: 'manual_preview',
+        duration_ms: Math.round(performance.now() - previewStartedAt),
+      })
+    }
 
     // Play the audio
     setTimeout(() => {

packages/stage-ui/src/components/scenes/Stage.vue

@@ -26,6 +26,7 @@ import { storeToRefs } from 'pinia'
 import { computed, onMounted, onUnmounted, ref, watch } from 'vue'
 
 import { useSettingsLive2d } from '../../../../stage-ui-live2d/src/composables/live2d/live2d'
+import { useAnalytics } from '../../composables/use-analytics'
 import { useAuthProviderSync } from '../../composables/use-auth-provider-sync'
 import { useDuckDb } from '../../composables/use-duck-db'
 import { useIOTraceBridge } from '../../composables/use-io-trace-bridge'
@@ -155,6 +156,8 @@ const speechStore = useSpeechStore()
 const { ssmlEnabled, activeSpeechProvider, activeSpeechModel, activeSpeechVoice, pitch } = storeToRefs(speechStore)
 const activeCardId = computed(() => activeCard.value?.name ?? 'default')
 const speechRuntimeStore = useSpeechRuntimeStore()
+const { trackOfficialTtsAutoEnabled } = useAnalytics()
+let officialAutoTtsTrackedForTurn = false
 const backgroundStore = useBackgroundStore()
 const { activeBackgroundUrl } = storeToRefs(backgroundStore)
 
@@ -315,6 +318,11 @@ async function playFunction(item: Parameters<Parameters<typeof createPlaybackMan
 
     try {
       source.start(0)
+      if (item.intentId.startsWith('stream-')) {
+        const model = resolveStreamingSessionModel()
+        if (model)
+          trackOfficialAutoTtsForTurn(model)
+      }
     }
     catch {
       stopPlayback()
@@ -337,6 +345,24 @@ function resolveStageVoiceType(): 'official_selected' | 'custom_configured' {
   return activeSpeechProvider.value === OFFICIAL_SPEECH_PROVIDER_ID || activeSpeechProvider.value === OFFICIAL_SPEECH_STREAMING_PROVIDER_ID ? 'official_selected' : 'custom_configured'
 }
 
+/**
+ * Tracks official auto-TTS once per assistant turn when chat audio is actually used.
+ */
+function trackOfficialAutoTtsForTurn(modelId: string) {
+  if (officialAutoTtsTrackedForTurn)
+    return
+  if (activeSpeechProvider.value !== OFFICIAL_SPEECH_PROVIDER_ID && activeSpeechProvider.value !== OFFICIAL_SPEECH_STREAMING_PROVIDER_ID)
+    return
+
+  officialAutoTtsTrackedForTurn = true
+  trackOfficialTtsAutoEnabled({
+    tts_provider_id: activeSpeechProvider.value,
+    tts_model_id: modelId,
+    source: 'chat_auto_tts',
+    enabled: true,
+  })
+}
+
 const speechPipeline = createSpeechPipeline<AudioBuffer>({
   tts: async (request, signal) => {
     if (signal.aborted)
@@ -459,6 +485,7 @@ const speechPipeline = createSpeechPipeline<AudioBuffer>({
         return null
 
       const audioBuffer = await audioContext.decodeAudioData(res)
+      trackOfficialAutoTtsForTurn(model)
       return audioBuffer
     }
     catch (err) {
@@ -631,6 +658,17 @@ function stopSpeechOutput(reason: string) {
   resetAssistantSpeechSurface(reason)
 }
 
+/**
+ * Resolves the official streaming TTS model for the current Stage session.
+ */
+function resolveStreamingSessionModel(): string | null {
+  const activeModel = activeSpeechModel.value as string | undefined
+  const sessionModel = activeModel?.includes('/') ? activeModel : getDefaultStreamingModel()
+  if (!sessionModel?.includes('/'))
+    return null
+  return sessionModel
+}
+
 function buildStreamingSnapshot(): StreamingSessionSnapshot | null {
   // Snapshotted once per session, so a mid-session provider/voice swap
   // does not corrupt an in-flight session — the watcher below detects
@@ -647,9 +685,8 @@ function buildStreamingSnapshot(): StreamingSessionSnapshot | null {
   // a provider switch) must NOT reach the bridge, so fall back to the
   // server-curated default instead of a hardcoded id. Returns null (segmenter
   // fallback) when neither resolves, rather than guessing a resource id.
-  const activeModel = activeSpeechModel.value as string | undefined
-  const sessionModel = activeModel?.includes('/') ? activeModel : getDefaultStreamingModel()
-  if (!sessionModel?.includes('/'))
+  const sessionModel = resolveStreamingSessionModel()
+  if (!sessionModel)
     return null
   const apiResourceId = sessionModel.split('/', 2)[1]
   // TTS 2.0 / ICL 2.0 ship subtitles asynchronously relative to audio
@@ -727,6 +764,7 @@ watch(latestStopRequest, (request) => {
 })
 
 chatHookCleanups.push(onBeforeMessageComposed(async () => {
+  officialAutoTtsTrackedForTurn = false
   playbackManager.stopAll('new-message')
 
   setupAnalyser()

packages/stage-ui/src/composables/use-analytics.test.ts

@@ -256,6 +256,131 @@ describe('useAnalytics conversation product events', () => {
     })
   })
 
+  /**
+   * @example
+   * analytics.trackOfficialProviderSelected({ provider_id: 'official-provider', provider_mode: 'official', source: 'default_auto', auto_selected: true })
+   * expect(posthog.capture).toHaveBeenCalledWith('official_provider_selected', expect.objectContaining({ auto_selected: true }))
+   */
+  it('emits official provider and second-turn operating events with bounded fields', () => {
+    const analytics = useAnalytics()
+
+    analytics.trackOfficialProviderSelected({
+      provider_id: 'official-provider',
+      provider_mode: 'official',
+      source: 'default_auto',
+      auto_selected: true,
+      model_id: 'chat-auto',
+    })
+    analytics.trackSecondTurnStarted({
+      provider_id: 'official-provider',
+      provider_mode: 'official',
+      model_id: 'chat-auto',
+      source: 'text',
+      turn_index: 2,
+    })
+
+    expect(analyticsMocks.posthogCaptureMock).toHaveBeenNthCalledWith(1, 'official_provider_selected', {
+      surface: 'web',
+      provider_id: 'official-provider',
+      provider_mode: 'official',
+      source: 'default_auto',
+      auto_selected: true,
+      model_id: 'chat-auto',
+    })
+    expect(analyticsMocks.posthogCaptureMock).toHaveBeenNthCalledWith(2, 'second_turn_started', {
+      surface: 'web',
+      provider_id: 'official-provider',
+      provider_mode: 'official',
+      model_id: 'chat-auto',
+      source: 'text',
+      turn_index: 2,
+    })
+  })
+
+  /**
+   * @example
+   * analytics.trackOfficialTtsExposed({ source: 'post_first_chat', tts_provider_id: 'official-provider-speech', tts_model_id: 'stepfun/tts' })
+   * expect(posthog.capture).toHaveBeenCalledWith('official_tts_exposed', expect.objectContaining({ source: 'post_first_chat' }))
+   */
+  it('emits official TTS activation funnel events', () => {
+    const analytics = useAnalytics()
+
+    analytics.trackOfficialTtsExposed({
+      tts_provider_id: 'official-provider-speech',
+      tts_model_id: 'stepfun/tts',
+      source: 'post_first_chat',
+    })
+    analytics.trackOfficialTtsPreviewStarted({
+      tts_provider_id: 'official-provider-speech',
+      tts_model_id: 'stepfun/tts',
+      voice_id: 'longxiaochun_v2',
+      voice_type: 'official_selected',
+      source: 'manual_preview',
+    })
+    analytics.trackOfficialTtsPreviewSucceeded({
+      tts_provider_id: 'official-provider-speech',
+      tts_model_id: 'stepfun/tts',
+      voice_id: 'longxiaochun_v2',
+      voice_type: 'official_selected',
+      source: 'manual_preview',
+      duration_ms: 320,
+    })
+    analytics.trackOfficialTtsAutoEnabled({
+      tts_provider_id: 'official-provider-speech',
+      tts_model_id: 'stepfun/tts',
+      source: 'settings',
+      enabled: true,
+    })
+
+    expect(analyticsMocks.posthogCaptureMock).toHaveBeenNthCalledWith(1, 'official_tts_exposed', {
+      surface: 'web',
+      tts_provider_id: 'official-provider-speech',
+      tts_model_id: 'stepfun/tts',
+      source: 'post_first_chat',
+    })
+    expect(analyticsMocks.posthogCaptureMock).toHaveBeenNthCalledWith(2, 'official_tts_preview_started', {
+      surface: 'web',
+      tts_provider_id: 'official-provider-speech',
+      tts_model_id: 'stepfun/tts',
+      voice_id: 'longxiaochun_v2',
+      voice_type: 'official_selected',
+      source: 'manual_preview',
+    })
+    expect(analyticsMocks.posthogCaptureMock).toHaveBeenNthCalledWith(3, 'official_tts_preview_succeeded', {
+      surface: 'web',
+      tts_provider_id: 'official-provider-speech',
+      tts_model_id: 'stepfun/tts',
+      voice_id: 'longxiaochun_v2',
+      voice_type: 'official_selected',
+      source: 'manual_preview',
+      duration_ms: 320,
+    })
+    expect(analyticsMocks.posthogCaptureMock).toHaveBeenNthCalledWith(4, 'official_tts_auto_enabled', {
+      surface: 'web',
+      tts_provider_id: 'official-provider-speech',
+      tts_model_id: 'stepfun/tts',
+      source: 'settings',
+      enabled: true,
+    })
+  })
+
+  it('emits paywall exposure with balance bucket for monetization funnels', () => {
+    const analytics = useAnalytics()
+
+    analytics.trackPaywallSeen({
+      surface: 'settings_flux',
+      reason: 'manual_topup',
+      flux_balance_bucket: '1_100',
+    })
+
+    expect(analyticsMocks.posthogCaptureMock).toHaveBeenCalledWith('paywall_seen', {
+      surface: 'settings_flux',
+      app_surface: 'web',
+      reason: 'manual_topup',
+      flux_balance_bucket: '1_100',
+    })
+  })
+
   /**
    * @example
    * analytics.trackMicrophonePermissionDenied({ stt_provider_id: 'browser-web-speech-api' })

packages/stage-ui/src/composables/use-analytics.ts

@@ -25,6 +25,9 @@ export type ChatActivationFailureStage = 'provider_config' | 'model_list' | 'mes
 export type ProviderConfigStep = 'settings_auto_validate' | 'manual_chat_ping' | 'onboarding_validate'
 export type VoiceType = 'official_default' | 'official_selected' | 'custom_configured' | 'voice_pack' | 'unknown'
 export type VoiceAnalyticsSource = 'settings' | 'onboarding' | 'chat_auto_tts' | 'manual_preview'
+export type OfficialProviderSelectionSource = 'settings' | 'onboarding' | 'default_auto'
+export type OfficialTtsExposureSource = 'settings' | 'onboarding' | 'post_first_chat' | 'chat_controls'
+export type FluxBalanceBucket = 'zero' | '1_100' | '101_1000' | '1001_10000' | '10000_plus' | 'unknown'
 export type FeedbackSource = 'app' | 'discord' | 'qq' | 'github' | 'email' | 'other'
 export type FeedbackCategory = 'provider_config' | 'model_list' | 'chat_activation' | 'tts' | 'voice_input' | 'performance' | 'payment' | 'ui_ux' | 'crash' | 'update' | 'live2d' | 'desktop_window' | 'mobile' | 'unknown'
 export type FeedbackSeverity = 'blocker' | 'major' | 'minor' | 'suggestion'
@@ -47,6 +50,12 @@ interface TtsVoiceBaseProperties {
   source: VoiceAnalyticsSource
 }
 
+interface OfficialTtsBaseProperties {
+  tts_provider_id: string
+  tts_model_id: string
+  source: OfficialTtsExposureSource
+}
+
 interface VoiceInputBaseProperties {
   stt_provider_id: string
   duration_ms?: number
@@ -191,6 +200,21 @@ export function useAnalytics() {
     )
   }
 
+  function trackPaywallSeen(properties: {
+    surface: string
+    reason: 'manual_topup' | 'insufficient_balance' | 'checkout_recovery' | 'unknown'
+    flux_balance_bucket: FluxBalanceBucket
+  }) {
+    if (!canCapture())
+      return
+    posthog.capture('paywall_seen', {
+      surface: properties.surface,
+      app_surface: getConversationAnalyticsSurface(),
+      reason: properties.reason,
+      flux_balance_bucket: properties.flux_balance_bucket,
+    })
+  }
+
   /** Activation funnel — step 1. */
   function trackSignup(method: 'email' | 'google' | 'github' | string) {
     if (!canCapture())
@@ -370,6 +394,21 @@ export function useAnalytics() {
     })
   }
 
+  function trackOfficialProviderSelected(properties: {
+    provider_id: string
+    provider_mode: ProviderMode
+    source: OfficialProviderSelectionSource
+    auto_selected: boolean
+    model_id?: string
+  }) {
+    if (!canCapture())
+      return
+    posthog.capture('official_provider_selected', {
+      ...properties,
+      surface: getConversationAnalyticsSurface(),
+    })
+  }
+
   function trackMessageSent(properties: ConversationBaseProperties & {
     message_id?: string
     message_index?: number
@@ -409,6 +448,15 @@ export function useAnalytics() {
     })
   }
 
+  function trackSecondTurnStarted(properties: ChatActivationBaseProperties & { turn_index: number }) {
+    if (!canCapture())
+      return
+    posthog.capture('second_turn_started', {
+      ...properties,
+      surface: getConversationAnalyticsSurface(),
+    })
+  }
+
   function trackModelListLoaded(properties: {
     provider_id: string
     provider_mode: ProviderMode
@@ -794,6 +842,15 @@ export function useAnalytics() {
     })
   }
 
+  function trackOfficialTtsExposed(properties: OfficialTtsBaseProperties) {
+    if (!canCapture())
+      return
+    posthog.capture('official_tts_exposed', {
+      ...properties,
+      surface: getConversationAnalyticsSurface(),
+    })
+  }
+
   function trackPresetUsed(properties: {
     preset_id: string
     preset_type: 'character' | 'stage_model' | 'voice' | 'background' | 'unknown'
@@ -807,6 +864,20 @@ export function useAnalytics() {
     })
   }
 
+  function trackOfficialTtsPreviewStarted(properties: Omit<TtsVoiceBaseProperties, 'source'> & {
+    voice_id: string
+    voice_type: VoiceType
+    voice_pack_id?: string
+    source: Extract<VoiceAnalyticsSource, 'manual_preview'>
+  }) {
+    if (!canCapture())
+      return
+    posthog.capture('official_tts_preview_started', {
+      ...properties,
+      surface: getConversationAnalyticsSurface(),
+    })
+  }
+
   function trackModelChanged(properties: {
     from_model?: string
     to_model: string
@@ -821,6 +892,21 @@ export function useAnalytics() {
     })
   }
 
+  function trackOfficialTtsPreviewSucceeded(properties: Omit<TtsVoiceBaseProperties, 'source'> & {
+    voice_id: string
+    voice_type: VoiceType
+    voice_pack_id?: string
+    source: Extract<VoiceAnalyticsSource, 'manual_preview'>
+    duration_ms: number
+  }) {
+    if (!canCapture())
+      return
+    posthog.capture('official_tts_preview_succeeded', {
+      ...properties,
+      surface: getConversationAnalyticsSurface(),
+    })
+  }
+
   function trackProviderSwitched(properties: {
     from_provider?: string
     to_provider: string
@@ -863,6 +949,18 @@ export function useAnalytics() {
     })
   }
 
+  function trackOfficialTtsAutoEnabled(properties: Omit<TtsVoiceBaseProperties, 'source'> & {
+    source: Extract<VoiceAnalyticsSource, 'settings' | 'chat_auto_tts'>
+    enabled: boolean
+  }) {
+    if (!canCapture())
+      return
+    posthog.capture('official_tts_auto_enabled', {
+      ...properties,
+      surface: getConversationAnalyticsSurface(),
+    })
+  }
+
   // ─── Autonomous LLM path (artistry-autonomous bypasses chat orchestrator) ─
 
   function trackAutonomousGenerateText(properties: { model: string, reason?: string }) {
@@ -983,6 +1081,7 @@ export function useAnalytics() {
     trackPricingViewed,
     trackPlanSelected,
     trackCheckoutStarted,
+    trackPaywallSeen,
     trackSignup,
     trackSignupCompleted,
     trackOnboardingStarted,
@@ -1005,6 +1104,8 @@ export function useAnalytics() {
     trackChatActivationStarted,
     trackChatActivationSucceeded,
     trackChatActivationFailed,
+    trackOfficialProviderSelected,
+    trackSecondTurnStarted,
     trackModelListLoaded,
     trackModelListFailed,
     trackProviderConfigStarted,
@@ -1049,6 +1150,10 @@ export function useAnalytics() {
     trackProviderSwitched,
     trackSettingsChanged,
     trackSupportContacted,
+    trackOfficialTtsExposed,
+    trackOfficialTtsPreviewStarted,
+    trackOfficialTtsPreviewSucceeded,
+    trackOfficialTtsAutoEnabled,
 
     trackAutonomousGenerateText,
 

packages/stage-ui/src/composables/use-auth-provider-sync.ts

@@ -8,6 +8,7 @@ import { useHearingStore } from '../stores/modules/hearing'
 import { useSpeechStore } from '../stores/modules/speech'
 import { useVisionStore } from '../stores/modules/vision'
 import { useProvidersStore } from '../stores/providers'
+import { useAnalytics } from './use-analytics'
 
 /**
  * Provider IDs to auto-activate on sign-in.
@@ -41,6 +42,7 @@ export function useAuthProviderSync() {
   const visionStore = useVisionStore()
   const speechStore = useSpeechStore()
   const hearingStore = useHearingStore()
+  const { trackOfficialProviderSelected } = useAnalytics()
 
   // Track whether the sync has already fired in this session to avoid
   // re-running on every page navigation (onAuthenticated fires immediately
@@ -68,6 +70,13 @@ export function useAuthProviderSync() {
           if (!consciousnessStore.activeProvider) {
             consciousnessStore.activeProvider = id
             consciousnessStore.activeModel = 'auto'
+            trackOfficialProviderSelected({
+              provider_id: id,
+              provider_mode: 'official',
+              source: 'default_auto',
+              auto_selected: true,
+              model_id: 'auto',
+            })
           }
           break
         case 'vision':

packages/stage-ui/src/stores/chat.contract.test.ts

@@ -39,6 +39,7 @@ const ioTracerMocks = vi.hoisted(() => {
 
 const llmStreamMock = vi.fn()
 const trackFirstMessageMock = vi.fn()
+const trackSecondTurnStartedMock = vi.fn()
 const ingestContextMessageMock = vi.fn()
 const getContextsSnapshotMock = vi.fn()
 const createMinecraftContextMock = vi.fn()
@@ -47,6 +48,8 @@ const forkSessionMock = vi.fn()
 const ensureSessionMock = vi.fn()
 
 const activeSessionIdRef = ref('session-1')
+const activeProviderRef = ref('mock-provider')
+const activeModelRef = ref('gpt-test')
 const streamingMessageRef = ref<any>({ role: 'assistant', content: '', slices: [], tool_results: [] })
 const sessionMessages: Record<string, any[]> = {}
 let currentGeneration = 1
@@ -62,14 +65,20 @@ vi.mock('pinia', async () => {
 vi.mock('../composables', () => ({
   useAnalytics: () => ({
     trackFirstMessage: trackFirstMessageMock,
+    trackChatFailed: vi.fn(),
+    trackChatStarted: vi.fn(),
     trackMessageSendStarted: vi.fn(),
+    trackMessageSent: vi.fn(),
     trackLlmRequestStarted: vi.fn(),
     trackLlmFirstToken: vi.fn(),
     trackAssistantResponseRendered: vi.fn(),
+    trackAssistantResponseCompleted: vi.fn(),
     trackMessageRound: vi.fn(),
+    trackFeatureUsed: vi.fn(),
     trackChatActivationStarted: vi.fn(),
     trackChatActivationSucceeded: vi.fn(),
     trackChatActivationFailed: vi.fn(),
+    trackSecondTurnStarted: trackSecondTurnStartedMock,
   }),
 }))
 
@@ -131,7 +140,8 @@ vi.mock('./llm-toolset-prompts', () => ({
 
 vi.mock('./modules/consciousness', () => ({
   useConsciousnessStore: () => ({
-    activeProvider: ref('mock-provider'),
+    activeModel: activeModelRef,
+    activeProvider: activeProviderRef,
   }),
 }))
 
@@ -156,6 +166,7 @@ describe('chat orchestrator contract', () => {
     setActivePinia(createPinia())
     llmStreamMock.mockReset()
     trackFirstMessageMock.mockReset()
+    trackSecondTurnStartedMock.mockReset()
     ingestContextMessageMock.mockReset()
     getContextsSnapshotMock.mockReset()
     getContextsSnapshotMock.mockReturnValue({})
@@ -168,6 +179,7 @@ describe('chat orchestrator contract', () => {
     ioTracerMocks.spans.length = 0
     ioTracerMocks.startSpanMock.mockClear()
     activeSessionIdRef.value = 'session-1'
+    activeProviderRef.value = 'mock-provider'
     streamingMessageRef.value = { role: 'assistant', content: '', slices: [], tool_results: [] }
     currentGeneration = 1
 
@@ -178,6 +190,34 @@ describe('chat orchestrator contract', () => {
     sessionMessages['session-1'] = [{ role: 'system', content: 'system prompt', createdAt: 1, id: 'system' }]
   })
 
+  it('emits second turn analytics from chat sends', async () => {
+    activeProviderRef.value = 'official-provider'
+    llmStreamMock.mockImplementation(async (_model: string, _chatProvider: ChatProvider, _messages: Message[], options: any) => {
+      await options.onStreamEvent({ type: 'text-delta', text: 'ok' })
+      await options.onStreamEvent({ type: 'finish', finishReason: 'stop' })
+    })
+
+    const store = useChatOrchestratorStore()
+
+    await store.ingest('first turn', {
+      model: 'chat-auto',
+      chatProvider: provider,
+    })
+    await store.ingest('second turn', {
+      model: 'chat-auto',
+      chatProvider: provider,
+    })
+
+    expect(trackSecondTurnStartedMock).toHaveBeenCalledTimes(1)
+    expect(trackSecondTurnStartedMock).toHaveBeenCalledWith({
+      provider_id: 'official-provider',
+      provider_mode: 'official',
+      model_id: 'chat-auto',
+      source: 'text',
+      turn_index: 2,
+    })
+  })
+
   it('keeps hook order and composes context prompt after system message', async () => {
     const contextsSnapshot = {
       'system:weather': [

packages/stage-ui/src/stores/chat.ts

@@ -64,6 +64,7 @@ export const useChatOrchestratorStore = defineStore('chat-orchestrator', () => {
     trackChatActivationStarted,
     trackChatActivationSucceeded,
     trackChatActivationFailed,
+    trackSecondTurnStarted,
   } = useAnalytics()
 
   const chatSession = useChatSessionStore()
@@ -225,12 +226,18 @@ export const useChatOrchestratorStore = defineStore('chat-orchestrator', () => {
       has_voice: hasVoice,
       model,
     }),
-    onChatActivationStarted: ({ model, provider, source }) => trackChatActivationStarted({
-      provider_mode: providerMode(provider),
-      provider_id: provider || 'unknown',
-      model_id: model || 'unknown',
-      source,
-    }),
+    onChatActivationStarted: ({ model, provider, source }) => {
+      const mode = providerMode(provider)
+      const providerId = provider || 'unknown'
+      const modelId = model || 'unknown'
+
+      trackChatActivationStarted({
+        provider_mode: mode,
+        provider_id: providerId,
+        model_id: modelId,
+        source,
+      })
+    },
     onChatActivationSucceeded: ({ model, provider, durationMs, source }) => trackChatActivationSucceeded({
       provider_mode: providerMode(provider),
       provider_id: provider || 'unknown',
@@ -258,7 +265,7 @@ export const useChatOrchestratorStore = defineStore('chat-orchestrator', () => {
     },
     onLifecycle: record => contextObservability.recordLifecycle(record),
     onPromptProjection: payload => contextObservability.capturePromptProjection(payload),
-    onUserMessageAppended: ({ sessionId, message, messageText }) => {
+    onUserMessageAppended: ({ sessionId, message, messageText, source, model, provider, turnIndex }) => {
       trackMessageSent({
         conversation_id: sessionId,
         provider_type: providerMode(activeProvider.value),
@@ -276,6 +283,16 @@ export const useChatOrchestratorStore = defineStore('chat-orchestrator', () => {
         entry: 'chat',
         success: true,
       })
+      if (turnIndex === 2) {
+        trackSecondTurnStarted({
+          provider_mode: providerMode(provider),
+          provider_id: provider || 'unknown',
+          model_id: model || 'unknown',
+          source,
+          turn_index: turnIndex,
+        })
+      }
+
       if (isCloudSyncableMessage(message)) {
         void chatSession.pushMessageToCloud(sessionId, {
           id: message.id,