diff --git a/CHANGELOG.md b/CHANGELOG.md
index dbef0a72951..ef0e4629b90 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -21,6 +21,7 @@ Docs: https://docs.openclaw.ai
 - Agents/compaction: surface safeguard-specific cancel reasons and relabel benign manual `/compact` no-op cases as skipped instead of failed. (#51072) Thanks @afurm.
 - Plugins/CLI backends: move bundled Claude CLI, Codex CLI, and Gemini CLI inference defaults onto the plugin surface, add bundled Gemini CLI backend support, and replace `gateway run --claude-cli-logs` with generic `--cli-backend-logs` while keeping the old flag as a compatibility alias.
 - Plugins/startup: auto-load bundled provider and CLI-backend plugins from explicit config refs, so bundled Claude CLI, Codex CLI, and Gemini CLI message-provider setups no longer need manual `plugins.allow` entries.
+- Config/TTS: auto-migrate legacy speech config on normal reads and secret resolution, keep legacy diagnostics for Doctor, and remove regular-mode runtime fallback for old bundled `tts.<provider>` API-key shapes.
 
 ### Fixes
 
diff --git a/docs/tools/tts.md b/docs/tools/tts.md
index f87c47eb725..27aea56026b 100644
--- a/docs/tools/tts.md
+++ b/docs/tools/tts.md
@@ -15,7 +15,7 @@ It works anywhere OpenClaw can send audio.
 ## Supported services
 
 - **ElevenLabs** (primary or fallback provider)
-- **Microsoft** (primary or fallback provider; current bundled implementation uses `node-edge-tts`, default when no API keys)
+- **Microsoft** (primary or fallback provider; current bundled implementation uses `node-edge-tts`)
 - **OpenAI** (primary or fallback provider; also used for summaries)
 
 ### Microsoft speech notes
@@ -38,9 +38,7 @@ If you want OpenAI or ElevenLabs:
 - `ELEVENLABS_API_KEY` (or `XI_API_KEY`)
 - `OPENAI_API_KEY`
 
-Microsoft speech does **not** require an API key. If no API keys are found,
-OpenClaw defaults to Microsoft (unless disabled via
-`messages.tts.microsoft.enabled=false` or `messages.tts.edge.enabled=false`).
+Microsoft speech does **not** require an API key.
 
 If multiple providers are configured, the selected provider is used first and the others are fallback options.
 Auto-summary uses the configured `summaryModel` (or `agents.defaults.model.primary`),
@@ -60,8 +58,8 @@ so that provider must also be authenticated if you enable summaries.
 No. Auto‑TTS is **off** by default. Enable it in config with
 `messages.tts.auto` or per session with `/tts always` (alias: `/tts on`).
 
-Microsoft speech **is** enabled by default once TTS is on, and is used automatically
-when no OpenAI or ElevenLabs API keys are available.
+When `messages.tts.provider` is unset, OpenClaw picks the first configured
+speech provider in registry auto-select order.
 
 ## Config
 
@@ -93,26 +91,28 @@ Full schema is in [Gateway configuration](/gateway/configuration).
       modelOverrides: {
         enabled: true,
       },
-      openai: {
-        apiKey: "openai_api_key",
-        baseUrl: "https://api.openai.com/v1",
-        model: "gpt-4o-mini-tts",
-        voice: "alloy",
-      },
-      elevenlabs: {
-        apiKey: "elevenlabs_api_key",
-        baseUrl: "https://api.elevenlabs.io",
-        voiceId: "voice_id",
-        modelId: "eleven_multilingual_v2",
-        seed: 42,
-        applyTextNormalization: "auto",
-        languageCode: "en",
-        voiceSettings: {
-          stability: 0.5,
-          similarityBoost: 0.75,
-          style: 0.0,
-          useSpeakerBoost: true,
-          speed: 1.0,
+      providers: {
+        openai: {
+          apiKey: "openai_api_key",
+          baseUrl: "https://api.openai.com/v1",
+          model: "gpt-4o-mini-tts",
+          voice: "alloy",
+        },
+        elevenlabs: {
+          apiKey: "elevenlabs_api_key",
+          baseUrl: "https://api.elevenlabs.io",
+          voiceId: "voice_id",
+          modelId: "eleven_multilingual_v2",
+          seed: 42,
+          applyTextNormalization: "auto",
+          languageCode: "en",
+          voiceSettings: {
+            stability: 0.5,
+            similarityBoost: 0.75,
+            style: 0.0,
+            useSpeakerBoost: true,
+            speed: 1.0,
+          },
         },
       },
     },
@@ -128,13 +128,15 @@ Full schema is in [Gateway configuration](/gateway/configuration).
     tts: {
       auto: "always",
       provider: "microsoft",
-      microsoft: {
-        enabled: true,
-        voice: "en-US-MichelleNeural",
-        lang: "en-US",
-        outputFormat: "audio-24khz-48kbitrate-mono-mp3",
-        rate: "+10%",
-        pitch: "-5%",
+      providers: {
+        microsoft: {
+          enabled: true,
+          voice: "en-US-MichelleNeural",
+          lang: "en-US",
+          outputFormat: "audio-24khz-48kbitrate-mono-mp3",
+          rate: "+10%",
+          pitch: "-5%",
+        },
       },
     },
   },
@@ -147,8 +149,10 @@ Full schema is in [Gateway configuration](/gateway/configuration).
 {
   messages: {
     tts: {
-      microsoft: {
-        enabled: false,
+      providers: {
+        microsoft: {
+          enabled: false,
+        },
       },
     },
   },
@@ -208,37 +212,37 @@ Then run:
 - `enabled`: legacy toggle (doctor migrates this to `auto`).
 - `mode`: `"final"` (default) or `"all"` (includes tool/block replies).
 - `provider`: speech provider id such as `"elevenlabs"`, `"microsoft"`, or `"openai"` (fallback is automatic).
-- If `provider` is **unset**, OpenClaw prefers `openai` (if key), then `elevenlabs` (if key),
-  otherwise `microsoft`.
+- If `provider` is **unset**, OpenClaw uses the first configured speech provider in registry auto-select order.
 - Legacy `provider: "edge"` still works and is normalized to `microsoft`.
 - `summaryModel`: optional cheap model for auto-summary; defaults to `agents.defaults.model.primary`.
   - Accepts `provider/model` or a configured model alias.
 - `modelOverrides`: allow the model to emit TTS directives (on by default).
   - `allowProvider` defaults to `false` (provider switching is opt-in).
+- `providers.<id>`: provider-owned settings keyed by speech provider id.
 - `maxTextLength`: hard cap for TTS input (chars). `/tts audio` fails if exceeded.
 - `timeoutMs`: request timeout (ms).
 - `prefsPath`: override the local prefs JSON path (provider/limit/summary).
 - `apiKey` values fall back to env vars (`ELEVENLABS_API_KEY`/`XI_API_KEY`, `OPENAI_API_KEY`).
-- `elevenlabs.baseUrl`: override ElevenLabs API base URL.
-- `openai.baseUrl`: override the OpenAI TTS endpoint.
-  - Resolution order: `messages.tts.openai.baseUrl` -> `OPENAI_TTS_BASE_URL` -> `https://api.openai.com/v1`
+- `providers.elevenlabs.baseUrl`: override ElevenLabs API base URL.
+- `providers.openai.baseUrl`: override the OpenAI TTS endpoint.
+  - Resolution order: `messages.tts.providers.openai.baseUrl` -> `OPENAI_TTS_BASE_URL` -> `https://api.openai.com/v1`
   - Non-default values are treated as OpenAI-compatible TTS endpoints, so custom model and voice names are accepted.
-- `elevenlabs.voiceSettings`:
+- `providers.elevenlabs.voiceSettings`:
   - `stability`, `similarityBoost`, `style`: `0..1`
   - `useSpeakerBoost`: `true|false`
   - `speed`: `0.5..2.0` (1.0 = normal)
-- `elevenlabs.applyTextNormalization`: `auto|on|off`
-- `elevenlabs.languageCode`: 2-letter ISO 639-1 (e.g. `en`, `de`)
-- `elevenlabs.seed`: integer `0..4294967295` (best-effort determinism)
-- `microsoft.enabled`: allow Microsoft speech usage (default `true`; no API key).
-- `microsoft.voice`: Microsoft neural voice name (e.g. `en-US-MichelleNeural`).
-- `microsoft.lang`: language code (e.g. `en-US`).
-- `microsoft.outputFormat`: Microsoft output format (e.g. `audio-24khz-48kbitrate-mono-mp3`).
+- `providers.elevenlabs.applyTextNormalization`: `auto|on|off`
+- `providers.elevenlabs.languageCode`: 2-letter ISO 639-1 (e.g. `en`, `de`)
+- `providers.elevenlabs.seed`: integer `0..4294967295` (best-effort determinism)
+- `providers.microsoft.enabled`: allow Microsoft speech usage (default `true`; no API key).
+- `providers.microsoft.voice`: Microsoft neural voice name (e.g. `en-US-MichelleNeural`).
+- `providers.microsoft.lang`: language code (e.g. `en-US`).
+- `providers.microsoft.outputFormat`: Microsoft output format (e.g. `audio-24khz-48kbitrate-mono-mp3`).
   - See Microsoft Speech output formats for valid values; not all formats are supported by the bundled Edge-backed transport.
-- `microsoft.rate` / `microsoft.pitch` / `microsoft.volume`: percent strings (e.g. `+10%`, `-5%`).
-- `microsoft.saveSubtitles`: write JSON subtitles alongside the audio file.
-- `microsoft.proxy`: proxy URL for Microsoft speech requests.
-- `microsoft.timeoutMs`: request timeout override (ms).
+- `providers.microsoft.rate` / `providers.microsoft.pitch` / `providers.microsoft.volume`: percent strings (e.g. `+10%`, `-5%`).
+- `providers.microsoft.saveSubtitles`: write JSON subtitles alongside the audio file.
+- `providers.microsoft.proxy`: proxy URL for Microsoft speech requests.
+- `providers.microsoft.timeoutMs`: request timeout override (ms).
 - `edge.*`: legacy alias for the same Microsoft settings.
 
 ## Model-driven overrides (default on)
diff --git a/docs/tts.md b/docs/tts.md
index ffd56f8b02e..6c3cea7ba05 100644
--- a/docs/tts.md
+++ b/docs/tts.md
@@ -15,7 +15,7 @@ It works anywhere OpenClaw can send audio.
 ## Supported services
 
 - **ElevenLabs** (primary or fallback provider)
-- **Microsoft** (primary or fallback provider; current bundled implementation uses `node-edge-tts`, default when no API keys)
+- **Microsoft** (primary or fallback provider; current bundled implementation uses `node-edge-tts`)
 - **OpenAI** (primary or fallback provider; also used for summaries)
 
 ### Microsoft speech notes
@@ -38,9 +38,7 @@ If you want OpenAI or ElevenLabs:
 - `ELEVENLABS_API_KEY` (or `XI_API_KEY`)
 - `OPENAI_API_KEY`
 
-Microsoft speech does **not** require an API key. If no API keys are found,
-OpenClaw defaults to Microsoft (unless disabled via
-`messages.tts.microsoft.enabled=false` or `messages.tts.edge.enabled=false`).
+Microsoft speech does **not** require an API key.
 
 If multiple providers are configured, the selected provider is used first and the others are fallback options.
 Auto-summary uses the configured `summaryModel` (or `agents.defaults.model.primary`),
@@ -60,8 +58,8 @@ so that provider must also be authenticated if you enable summaries.
 No. Auto‑TTS is **off** by default. Enable it in config with
 `messages.tts.auto` or per session with `/tts always` (alias: `/tts on`).
 
-Microsoft speech **is** enabled by default once TTS is on, and is used automatically
-when no OpenAI or ElevenLabs API keys are available.
+When `messages.tts.provider` is unset, OpenClaw picks the first configured
+speech provider in registry auto-select order.
 
 ## Config
 
@@ -93,26 +91,28 @@ Full schema is in [Gateway configuration](/gateway/configuration).
       modelOverrides: {
         enabled: true,
       },
-      openai: {
-        apiKey: "openai_api_key",
-        baseUrl: "https://api.openai.com/v1",
-        model: "gpt-4o-mini-tts",
-        voice: "alloy",
-      },
-      elevenlabs: {
-        apiKey: "elevenlabs_api_key",
-        baseUrl: "https://api.elevenlabs.io",
-        voiceId: "voice_id",
-        modelId: "eleven_multilingual_v2",
-        seed: 42,
-        applyTextNormalization: "auto",
-        languageCode: "en",
-        voiceSettings: {
-          stability: 0.5,
-          similarityBoost: 0.75,
-          style: 0.0,
-          useSpeakerBoost: true,
-          speed: 1.0,
+      providers: {
+        openai: {
+          apiKey: "openai_api_key",
+          baseUrl: "https://api.openai.com/v1",
+          model: "gpt-4o-mini-tts",
+          voice: "alloy",
+        },
+        elevenlabs: {
+          apiKey: "elevenlabs_api_key",
+          baseUrl: "https://api.elevenlabs.io",
+          voiceId: "voice_id",
+          modelId: "eleven_multilingual_v2",
+          seed: 42,
+          applyTextNormalization: "auto",
+          languageCode: "en",
+          voiceSettings: {
+            stability: 0.5,
+            similarityBoost: 0.75,
+            style: 0.0,
+            useSpeakerBoost: true,
+            speed: 1.0,
+          },
         },
       },
     },
@@ -128,13 +128,15 @@ Full schema is in [Gateway configuration](/gateway/configuration).
     tts: {
       auto: "always",
       provider: "microsoft",
-      microsoft: {
-        enabled: true,
-        voice: "en-US-MichelleNeural",
-        lang: "en-US",
-        outputFormat: "audio-24khz-48kbitrate-mono-mp3",
-        rate: "+10%",
-        pitch: "-5%",
+      providers: {
+        microsoft: {
+          enabled: true,
+          voice: "en-US-MichelleNeural",
+          lang: "en-US",
+          outputFormat: "audio-24khz-48kbitrate-mono-mp3",
+          rate: "+10%",
+          pitch: "-5%",
+        },
       },
     },
   },
@@ -147,8 +149,10 @@ Full schema is in [Gateway configuration](/gateway/configuration).
 {
   messages: {
     tts: {
-      microsoft: {
-        enabled: false,
+      providers: {
+        microsoft: {
+          enabled: false,
+        },
       },
     },
   },
@@ -208,37 +212,37 @@ Then run:
 - `enabled`: legacy toggle (doctor migrates this to `auto`).
 - `mode`: `"final"` (default) or `"all"` (includes tool/block replies).
 - `provider`: speech provider id such as `"elevenlabs"`, `"microsoft"`, or `"openai"` (fallback is automatic).
-- If `provider` is **unset**, OpenClaw prefers `openai` (if key), then `elevenlabs` (if key),
-  otherwise `microsoft`.
+- If `provider` is **unset**, OpenClaw uses the first configured speech provider in registry auto-select order.
 - Legacy `provider: "edge"` still works and is normalized to `microsoft`.
 - `summaryModel`: optional cheap model for auto-summary; defaults to `agents.defaults.model.primary`.
   - Accepts `provider/model` or a configured model alias.
 - `modelOverrides`: allow the model to emit TTS directives (on by default).
   - `allowProvider` defaults to `false` (provider switching is opt-in).
+- `providers.<id>`: provider-owned settings keyed by speech provider id.
 - `maxTextLength`: hard cap for TTS input (chars). `/tts audio` fails if exceeded.
 - `timeoutMs`: request timeout (ms).
 - `prefsPath`: override the local prefs JSON path (provider/limit/summary).
 - `apiKey` values fall back to env vars (`ELEVENLABS_API_KEY`/`XI_API_KEY`, `OPENAI_API_KEY`).
-- `elevenlabs.baseUrl`: override ElevenLabs API base URL.
-- `openai.baseUrl`: override the OpenAI TTS endpoint.
-  - Resolution order: `messages.tts.openai.baseUrl` -> `OPENAI_TTS_BASE_URL` -> `https://api.openai.com/v1`
+- `providers.elevenlabs.baseUrl`: override ElevenLabs API base URL.
+- `providers.openai.baseUrl`: override the OpenAI TTS endpoint.
+  - Resolution order: `messages.tts.providers.openai.baseUrl` -> `OPENAI_TTS_BASE_URL` -> `https://api.openai.com/v1`
   - Non-default values are treated as OpenAI-compatible TTS endpoints, so custom model and voice names are accepted.
-- `elevenlabs.voiceSettings`:
+- `providers.elevenlabs.voiceSettings`:
   - `stability`, `similarityBoost`, `style`: `0..1`
   - `useSpeakerBoost`: `true|false`
   - `speed`: `0.5..2.0` (1.0 = normal)
-- `elevenlabs.applyTextNormalization`: `auto|on|off`
-- `elevenlabs.languageCode`: 2-letter ISO 639-1 (e.g. `en`, `de`)
-- `elevenlabs.seed`: integer `0..4294967295` (best-effort determinism)
-- `microsoft.enabled`: allow Microsoft speech usage (default `true`; no API key).
-- `microsoft.voice`: Microsoft neural voice name (e.g. `en-US-MichelleNeural`).
-- `microsoft.lang`: language code (e.g. `en-US`).
-- `microsoft.outputFormat`: Microsoft output format (e.g. `audio-24khz-48kbitrate-mono-mp3`).
+- `providers.elevenlabs.applyTextNormalization`: `auto|on|off`
+- `providers.elevenlabs.languageCode`: 2-letter ISO 639-1 (e.g. `en`, `de`)
+- `providers.elevenlabs.seed`: integer `0..4294967295` (best-effort determinism)
+- `providers.microsoft.enabled`: allow Microsoft speech usage (default `true`; no API key).
+- `providers.microsoft.voice`: Microsoft neural voice name (e.g. `en-US-MichelleNeural`).
+- `providers.microsoft.lang`: language code (e.g. `en-US`).
+- `providers.microsoft.outputFormat`: Microsoft output format (e.g. `audio-24khz-48kbitrate-mono-mp3`).
   - See Microsoft Speech output formats for valid values; not all formats are supported by the bundled Edge-backed transport.
-- `microsoft.rate` / `microsoft.pitch` / `microsoft.volume`: percent strings (e.g. `+10%`, `-5%`).
-- `microsoft.saveSubtitles`: write JSON subtitles alongside the audio file.
-- `microsoft.proxy`: proxy URL for Microsoft speech requests.
-- `microsoft.timeoutMs`: request timeout override (ms).
+- `providers.microsoft.rate` / `providers.microsoft.pitch` / `providers.microsoft.volume`: percent strings (e.g. `+10%`, `-5%`).
+- `providers.microsoft.saveSubtitles`: write JSON subtitles alongside the audio file.
+- `providers.microsoft.proxy`: proxy URL for Microsoft speech requests.
+- `providers.microsoft.timeoutMs`: request timeout override (ms).
 - `edge.*`: legacy alias for the same Microsoft settings.
 
 ## Model-driven overrides (default on)
diff --git a/src/commands/doctor-legacy-config.migrations.test.ts b/src/commands/doctor-legacy-config.migrations.test.ts
index d4de96b8367..869b58f64a7 100644
--- a/src/commands/doctor-legacy-config.migrations.test.ts
+++ b/src/commands/doctor-legacy-config.migrations.test.ts
@@ -523,7 +523,6 @@ describe("normalizeCompatibilityConfigValues", () => {
     });
 
     expect(res.config.talk).toEqual({
-      provider: "elevenlabs",
       providers: {
         elevenlabs: {
           voiceId: "voice-123",
@@ -545,9 +544,7 @@ describe("normalizeCompatibilityConfigValues", () => {
       interruptOnSpeech: false,
       silenceTimeoutMs: 1500,
     });
-    expect(res.changes).toEqual([
-      "Moved legacy talk flat fields → talk.provider/talk.providers.elevenlabs.",
-    ]);
+    expect(res.changes).toEqual(["Moved legacy talk flat fields → talk.providers.elevenlabs."]);
   });
 
   it("normalizes talk provider ids without overriding explicit provider config", () => {
diff --git a/src/commands/doctor-legacy-config.ts b/src/commands/doctor-legacy-config.ts
index 99d8783c76a..d489d354be0 100644
--- a/src/commands/doctor-legacy-config.ts
+++ b/src/commands/doctor-legacy-config.ts
@@ -11,7 +11,7 @@ import {
   resolveTelegramPreviewStreamMode,
 } from "../config/discord-preview-streaming.js";
 import { migrateLegacyWebSearchConfig } from "../config/legacy-web-search.js";
-import { DEFAULT_TALK_PROVIDER, normalizeTalkSection } from "../config/talk.js";
+import { LEGACY_TALK_PROVIDER_ID, normalizeTalkSection } from "../config/talk.js";
 import { DEFAULT_GOOGLE_API_BASE_URL } from "../infra/google-api-base-url.js";
 import { DEFAULT_ACCOUNT_ID } from "../routing/session-key.js";
 
@@ -651,9 +651,7 @@ export function normalizeCompatibilityConfigValues(cfg: OpenClawConfig): {
       return;
     }
 
-    changes.push(
-      `Moved legacy talk flat fields → talk.provider/talk.providers.${DEFAULT_TALK_PROVIDER}.`,
-    );
+    changes.push(`Moved legacy talk flat fields → talk.providers.${LEGACY_TALK_PROVIDER_ID}.`);
   };
 
   const normalizeLegacyCrossContextMessageConfig = () => {
diff --git a/src/config/config-misc.test.ts b/src/config/config-misc.test.ts
index f089e74015c..8de11cdb086 100644
--- a/src/config/config-misc.test.ts
+++ b/src/config/config-misc.test.ts
@@ -482,7 +482,7 @@ describe("config strict validation", () => {
 
       const snap = await readConfigFileSnapshot();
 
-      expect(snap.valid).toBe(false);
+      expect(snap.valid).toBe(true);
       expect(snap.legacyIssues).not.toHaveLength(0);
     });
   });
@@ -517,7 +517,7 @@ describe("config strict validation", () => {
       });
 
       const snap = await readConfigFileSnapshot();
-      expect(snap.valid).toBe(false);
+      expect(snap.valid).toBe(true);
       expect(snap.legacyIssues.some((issue) => issue.path === "gateway.bind")).toBe(true);
     });
   });
diff --git a/src/config/config.legacy-config-detection.accepts-imessage-dmpolicy.test.ts b/src/config/config.legacy-config-detection.accepts-imessage-dmpolicy.test.ts
index 89632bbc543..d21835ca910 100644
--- a/src/config/config.legacy-config-detection.accepts-imessage-dmpolicy.test.ts
+++ b/src/config/config.legacy-config-detection.accepts-imessage-dmpolicy.test.ts
@@ -68,7 +68,7 @@ function expectRoutingAllowFromLegacySnapshot(
   ctx: { snapshot: ConfigSnapshot; parsed: unknown },
   expectedAllowFrom: string[],
 ) {
-  expect(ctx.snapshot.valid).toBe(false);
+  expect(ctx.snapshot.valid).toBe(true);
   expect(ctx.snapshot.legacyIssues.some((issue) => issue.path === "routing.allowFrom")).toBe(true);
   const parsed = ctx.parsed as {
     routing?: { allowFrom?: string[] };
@@ -269,7 +269,7 @@ describe("legacy config detection", () => {
     await withSnapshotForConfig(
       { memorySearch: { provider: "local", fallback: "none" } },
       async (ctx) => {
-        expect(ctx.snapshot.valid).toBe(false);
+        expect(ctx.snapshot.valid).toBe(true);
         expect(ctx.snapshot.legacyIssues.some((issue) => issue.path === "memorySearch")).toBe(true);
       },
     );
@@ -278,14 +278,14 @@ describe("legacy config detection", () => {
     await withSnapshotForConfig(
       { heartbeat: { model: "anthropic/claude-3-5-haiku-20241022", every: "30m" } },
       async (ctx) => {
-        expect(ctx.snapshot.valid).toBe(false);
+        expect(ctx.snapshot.valid).toBe(true);
         expect(ctx.snapshot.legacyIssues.some((issue) => issue.path === "heartbeat")).toBe(true);
       },
     );
   });
   it("flags legacy provider sections in snapshot", async () => {
     await withSnapshotForConfig({ whatsapp: { allowFrom: ["+1555"] } }, async (ctx) => {
-      expect(ctx.snapshot.valid).toBe(false);
+      expect(ctx.snapshot.valid).toBe(true);
       expect(ctx.snapshot.legacyIssues.some((issue) => issue.path === "whatsapp")).toBe(true);
 
       const parsed = ctx.parsed as {
diff --git a/src/config/defaults.ts b/src/config/defaults.ts
index dcc66044737..d535ed0bc9a 100644
--- a/src/config/defaults.ts
+++ b/src/config/defaults.ts
@@ -3,7 +3,7 @@ import { normalizeProviderId, parseModelRef } from "../agents/model-selection.js
 import { DEFAULT_AGENT_MAX_CONCURRENT, DEFAULT_SUBAGENT_MAX_CONCURRENT } from "./agent-limits.js";
 import { resolveAgentModelPrimaryValue } from "./model-input.js";
 import {
-  DEFAULT_TALK_PROVIDER,
+  LEGACY_TALK_PROVIDER_ID,
   normalizeTalkConfig,
   resolveActiveTalkProviderConfig,
   resolveTalkApiKey,
@@ -204,7 +204,7 @@ export function applyTalkApiKey(config: OpenClawConfig): OpenClawConfig {
 
   const talk = normalized.talk;
   const active = resolveActiveTalkProviderConfig(talk);
-  if (active?.provider && active.provider !== DEFAULT_TALK_PROVIDER) {
+  if (!active || active.provider !== LEGACY_TALK_PROVIDER_ID) {
     return normalized;
   }
 
@@ -214,7 +214,7 @@ export function applyTalkApiKey(config: OpenClawConfig): OpenClawConfig {
     return normalized;
   }
 
-  const providerId = active?.provider ?? DEFAULT_TALK_PROVIDER;
+  const providerId = active.provider;
   const providers = { ...talk?.providers };
   const providerConfig = { ...providers[providerId], apiKey: resolved };
   providers[providerId] = providerConfig;
@@ -222,7 +222,6 @@ export function applyTalkApiKey(config: OpenClawConfig): OpenClawConfig {
   const nextTalk = {
     ...talk,
     apiKey: resolved,
-    provider: talk?.provider ?? providerId,
     providers,
   };
 
diff --git a/src/config/io.ts b/src/config/io.ts
index 953ff0747c8..c68739188b8 100644
--- a/src/config/io.ts
+++ b/src/config/io.ts
@@ -41,6 +41,7 @@ import {
   readConfigIncludeFileWithGuards,
   resolveConfigIncludes,
 } from "./includes.js";
+import { migrateLegacyConfig } from "./legacy-migrate.js";
 import { findLegacyConfigIssues } from "./legacy.js";
 import { applyMergePatch } from "./merge-patch.js";
 import { normalizeExecSafeBinProfilesInConfig } from "./normalize-exec-safe-bin.js";
@@ -1185,6 +1186,11 @@ type ConfigReadResolution = {
   envWarnings: EnvSubstitutionWarning[];
 };
 
+type LegacyMigrationResolution = {
+  effectiveConfigRaw: unknown;
+  sourceLegacyIssues: LegacyConfigIssue[];
+};
+
 function resolveConfigIncludesForRead(
   parsed: unknown,
   configPath: string,
@@ -1225,6 +1231,21 @@ function resolveConfigForRead(
   };
 }
 
+function resolveLegacyConfigForRead(
+  resolvedConfigRaw: unknown,
+  sourceRaw: unknown,
+): LegacyMigrationResolution {
+  const sourceLegacyIssues = findLegacyConfigIssues(resolvedConfigRaw, sourceRaw);
+  if (sourceLegacyIssues.length === 0) {
+    return { effectiveConfigRaw: resolvedConfigRaw, sourceLegacyIssues };
+  }
+  const migrated = migrateLegacyConfig(resolvedConfigRaw);
+  return {
+    effectiveConfigRaw: migrated.config ?? resolvedConfigRaw,
+    sourceLegacyIssues,
+  };
+}
+
 type ReadConfigFileSnapshotInternalResult = {
   snapshot: ConfigFileSnapshot;
   envSnapshotForRestore?: Record<string, string | undefined>;
@@ -1275,13 +1296,15 @@ export function createConfigIO(overrides: ConfigIoDeps = {}) {
         deps.env,
       );
       const resolvedConfig = readResolution.resolvedConfigRaw;
+      const legacyResolution = resolveLegacyConfigForRead(resolvedConfig, parsed);
+      const effectiveConfigRaw = legacyResolution.effectiveConfigRaw;
       for (const w of readResolution.envWarnings) {
         deps.logger.warn(
           `Config (${configPath}): missing env var "${w.varName}" at ${w.configPath} — feature using this value will be unavailable`,
         );
       }
-      warnOnConfigMiskeys(resolvedConfig, deps.logger);
-      if (typeof resolvedConfig !== "object" || resolvedConfig === null) {
+      warnOnConfigMiskeys(effectiveConfigRaw, deps.logger);
+      if (typeof effectiveConfigRaw !== "object" || effectiveConfigRaw === null) {
         observeLoadConfigSnapshot({
           path: configPath,
           exists: true,
@@ -1293,31 +1316,31 @@ export function createConfigIO(overrides: ConfigIoDeps = {}) {
           hash,
           issues: [],
           warnings: [],
-          legacyIssues: [],
+          legacyIssues: legacyResolution.sourceLegacyIssues,
         });
         return {};
       }
-      const preValidationDuplicates = findDuplicateAgentDirs(resolvedConfig as OpenClawConfig, {
+      const preValidationDuplicates = findDuplicateAgentDirs(effectiveConfigRaw as OpenClawConfig, {
         env: deps.env,
         homedir: deps.homedir,
       });
       if (preValidationDuplicates.length > 0) {
         throw new DuplicateAgentDirError(preValidationDuplicates);
       }
-      const validated = validateConfigObjectWithPlugins(resolvedConfig, { env: deps.env });
+      const validated = validateConfigObjectWithPlugins(effectiveConfigRaw, { env: deps.env });
       if (!validated.ok) {
         observeLoadConfigSnapshot({
           path: configPath,
           exists: true,
           raw,
           parsed,
-          resolved: coerceConfig(resolvedConfig),
+          resolved: coerceConfig(effectiveConfigRaw),
           valid: false,
-          config: coerceConfig(resolvedConfig),
+          config: coerceConfig(effectiveConfigRaw),
           hash,
           issues: validated.issues,
           warnings: validated.warnings,
-          legacyIssues: findLegacyConfigIssues(resolvedConfig, parsed),
+          legacyIssues: legacyResolution.sourceLegacyIssues,
         });
         const details = validated.issues
           .map(
@@ -1362,13 +1385,13 @@ export function createConfigIO(overrides: ConfigIoDeps = {}) {
         exists: true,
         raw,
         parsed,
-        resolved: coerceConfig(resolvedConfig),
+        resolved: coerceConfig(effectiveConfigRaw),
         valid: true,
         config: cfg,
         hash,
         issues: [],
         warnings: validated.warnings,
-        legacyIssues: findLegacyConfigIssues(resolvedConfig, parsed),
+        legacyIssues: legacyResolution.sourceLegacyIssues,
       });
 
       const duplicates = findDuplicateAgentDirs(cfg, {
@@ -1536,11 +1559,10 @@ export function createConfigIO(overrides: ConfigIoDeps = {}) {
       }));
 
       const resolvedConfigRaw = readResolution.resolvedConfigRaw;
-      // Detect legacy keys on resolved config, but only mark source-literal legacy
-      // entries (for auto-migration) when they are present in the parsed source.
-      const legacyIssues = findLegacyConfigIssues(resolvedConfigRaw, parsedRes.parsed);
+      const legacyResolution = resolveLegacyConfigForRead(resolvedConfigRaw, parsedRes.parsed);
+      const effectiveConfigRaw = legacyResolution.effectiveConfigRaw;
 
-      const validated = validateConfigObjectWithPlugins(resolvedConfigRaw, { env: deps.env });
+      const validated = validateConfigObjectWithPlugins(effectiveConfigRaw, { env: deps.env });
       if (!validated.ok) {
         return await finalizeReadConfigSnapshotInternalResult(deps, {
           snapshot: {
@@ -1548,13 +1570,13 @@ export function createConfigIO(overrides: ConfigIoDeps = {}) {
             exists: true,
             raw,
             parsed: parsedRes.parsed,
-            resolved: coerceConfig(resolvedConfigRaw),
+            resolved: coerceConfig(effectiveConfigRaw),
             valid: false,
-            config: coerceConfig(resolvedConfigRaw),
+            config: coerceConfig(effectiveConfigRaw),
             hash,
             issues: validated.issues,
             warnings: [...validated.warnings, ...envVarWarnings],
-            legacyIssues,
+            legacyIssues: legacyResolution.sourceLegacyIssues,
           },
         });
       }
@@ -1580,13 +1602,13 @@ export function createConfigIO(overrides: ConfigIoDeps = {}) {
           parsed: parsedRes.parsed,
           // Use resolvedConfigRaw (after $include and ${ENV} substitution but BEFORE runtime defaults)
           // for config set/unset operations (issue #6070)
-          resolved: coerceConfig(resolvedConfigRaw),
+          resolved: coerceConfig(effectiveConfigRaw),
           valid: true,
           config: snapshotConfig,
           hash,
           issues: [],
           warnings: [...validated.warnings, ...envVarWarnings],
-          legacyIssues,
+          legacyIssues: legacyResolution.sourceLegacyIssues,
         },
         envSnapshotForRestore: readResolution.envSnapshotForRestore,
       });
diff --git a/src/config/talk.normalize.test.ts b/src/config/talk.normalize.test.ts
index f2b1ddff1a1..a4221e5223c 100644
--- a/src/config/talk.normalize.test.ts
+++ b/src/config/talk.normalize.test.ts
@@ -36,7 +36,6 @@ describe("talk normalization", () => {
     });
 
     expect(normalized).toEqual({
-      provider: "elevenlabs",
       providers: {
         elevenlabs: {
           voiceId: "voice-123",
@@ -149,7 +148,7 @@ describe("talk normalization", () => {
         async (configPath) => {
           const io = createConfigIO({ configPath });
           const snapshot = await io.readConfigFileSnapshot();
-          expect(snapshot.config.talk?.provider).toBe("elevenlabs");
+          expect(snapshot.config.talk?.provider).toBeUndefined();
           expect(snapshot.config.talk?.providers?.elevenlabs?.voiceId).toBe("voice-123");
           expect(snapshot.config.talk?.providers?.elevenlabs?.apiKey).toBe(elevenLabsApiKey);
           expect(snapshot.config.talk?.apiKey).toBe(elevenLabsApiKey);
diff --git a/src/config/talk.ts b/src/config/talk.ts
index 32c4255a7a4..517e00ac252 100644
--- a/src/config/talk.ts
+++ b/src/config/talk.ts
@@ -16,7 +16,7 @@ type TalkApiKeyDeps = {
   path?: typeof path;
 };
 
-export const DEFAULT_TALK_PROVIDER = "elevenlabs";
+export const LEGACY_TALK_PROVIDER_ID = "elevenlabs";
 
 function isPlainObject(value: unknown): value is Record<string, unknown> {
   return typeof value === "object" && value !== null && !Array.isArray(value);
@@ -225,19 +225,13 @@ export function normalizeTalkSection(value: TalkConfig | undefined): TalkConfig
     }
     if (provider) {
       normalized.provider = provider;
-    } else if (providers) {
-      const ids = Object.keys(providers);
-      if (ids.length === 1) {
-        normalized.provider = ids[0];
-      }
     }
     return Object.keys(normalized).length > 0 ? normalized : undefined;
   }
 
   const legacyProviderConfig = legacyProviderConfigFromTalk(source);
   if (legacyProviderConfig) {
-    normalized.provider = DEFAULT_TALK_PROVIDER;
-    normalized.providers = { [DEFAULT_TALK_PROVIDER]: legacyProviderConfig };
+    normalized.providers = { [LEGACY_TALK_PROVIDER_ID]: legacyProviderConfig };
   }
   return Object.keys(normalized).length > 0 ? normalized : undefined;
 }
diff --git a/src/config/types.tts.ts b/src/config/types.tts.ts
index 934594a7253..bec252c9900 100644
--- a/src/config/types.tts.ts
+++ b/src/config/types.tts.ts
@@ -27,22 +27,8 @@ export type TtsModelOverrideConfig = {
 
 export type TtsProviderConfigMap = Record<string, Record<string, unknown>>;
 
-export type TtsConfig = {
-  /** Auto-TTS mode (preferred). */
-  auto?: TtsAutoMode;
-  /** Legacy: enable auto-TTS when `auto` is not set. */
-  enabled?: boolean;
-  /** Apply TTS to final replies only or to all replies (tool/block/final). */
-  mode?: TtsMode;
-  /** Primary TTS provider (fallbacks are automatic). */
-  provider?: TtsProvider;
-  /** Optional model override for TTS auto-summary (provider/model or alias). */
-  summaryModel?: string;
-  /** Allow the model to override TTS parameters. */
-  modelOverrides?: TtsModelOverrideConfig;
-  /** Provider-specific TTS settings keyed by speech provider id. */
-  providers?: TtsProviderConfigMap;
-  /** ElevenLabs configuration. */
+export type LegacyTtsConfigCompat = {
+  /** Legacy ElevenLabs configuration. Prefer providers.elevenlabs. */
   elevenlabs?: {
     apiKey?: SecretInput;
     baseUrl?: string;
@@ -59,7 +45,7 @@ export type TtsConfig = {
       speed?: number;
     };
   };
-  /** OpenAI configuration. */
+  /** Legacy OpenAI configuration. Prefer providers.openai. */
   openai?: {
     apiKey?: SecretInput;
     baseUrl?: string;
@@ -70,7 +56,7 @@ export type TtsConfig = {
     /** System-level instructions for the TTS model (gpt-4o-mini-tts only). */
     instructions?: string;
   };
-  /** Legacy alias for Microsoft speech configuration. */
+  /** Legacy alias for Microsoft speech configuration. Prefer providers.microsoft. */
   edge?: {
     /** Explicitly allow Microsoft speech usage (no API key required). */
     enabled?: boolean;
@@ -84,7 +70,7 @@ export type TtsConfig = {
     proxy?: string;
     timeoutMs?: number;
   };
-  /** Preferred alias for Microsoft speech configuration. */
+  /** Legacy Microsoft speech configuration. Prefer providers.microsoft. */
   microsoft?: {
     enabled?: boolean;
     voice?: string;
@@ -97,6 +83,23 @@ export type TtsConfig = {
     proxy?: string;
     timeoutMs?: number;
   };
+};
+
+export type TtsConfig = LegacyTtsConfigCompat & {
+  /** Auto-TTS mode (preferred). */
+  auto?: TtsAutoMode;
+  /** Legacy: enable auto-TTS when `auto` is not set. */
+  enabled?: boolean;
+  /** Apply TTS to final replies only or to all replies (tool/block/final). */
+  mode?: TtsMode;
+  /** Primary TTS provider (fallbacks are automatic). */
+  provider?: TtsProvider;
+  /** Optional model override for TTS auto-summary (provider/model or alias). */
+  summaryModel?: string;
+  /** Allow the model to override TTS parameters. */
+  modelOverrides?: TtsModelOverrideConfig;
+  /** Provider-specific TTS settings keyed by speech provider id. */
+  providers?: TtsProviderConfigMap;
   /** Optional path for local TTS user preferences JSON. */
   prefsPath?: string;
   /** Hard cap for text sent to TTS (chars). */
diff --git a/src/secrets/runtime-config-collectors-tts.ts b/src/secrets/runtime-config-collectors-tts.ts
index 1c93323a7f3..295ac9bf244 100644
--- a/src/secrets/runtime-config-collectors-tts.ts
+++ b/src/secrets/runtime-config-collectors-tts.ts
@@ -54,25 +54,4 @@ export function collectTtsApiKeyAssignments(params: {
     }
     return;
   }
-
-  // Legacy compatibility until migrated configs have been rewritten on disk.
-  const legacyProviders = ["elevenlabs", "openai"] as const;
-  for (const providerId of legacyProviders) {
-    const providerConfig = params.tts[providerId];
-    if (!isRecord(providerConfig)) {
-      continue;
-    }
-    collectSecretInputAssignment({
-      value: providerConfig.apiKey,
-      path: `${params.pathPrefix}.${providerId}.apiKey`,
-      expected: "string",
-      defaults: params.defaults,
-      context: params.context,
-      active: params.active,
-      inactiveReason: params.inactiveReason,
-      apply: (value) => {
-        providerConfig.apiKey = value;
-      },
-    });
-  }
 }
diff --git a/src/secrets/runtime.test.ts b/src/secrets/runtime.test.ts
index a93c216f1a2..6f892f195cf 100644
--- a/src/secrets/runtime.test.ts
+++ b/src/secrets/runtime.test.ts
@@ -1949,20 +1949,22 @@ describe("secrets runtime snapshot", () => {
       loadAuthStore: () => ({ version: 1, profiles: {} }),
     });
 
-    expect(snapshot.config.channels?.discord?.voice?.tts?.openai?.apiKey).toEqual({
+    expect(snapshot.config.channels?.discord?.voice?.tts?.providers?.openai?.apiKey).toEqual({
       source: "env",
       provider: "default",
       id: "MISSING_DISCORD_VOICE_TTS_OPENAI",
     });
-    expect(snapshot.config.channels?.discord?.accounts?.work?.voice?.tts?.openai?.apiKey).toEqual({
+    expect(
+      snapshot.config.channels?.discord?.accounts?.work?.voice?.tts?.providers?.openai?.apiKey,
+    ).toEqual({
       source: "env",
       provider: "default",
       id: "MISSING_DISCORD_WORK_VOICE_TTS_OPENAI",
     });
     expect(snapshot.warnings.map((warning) => warning.path)).toEqual(
       expect.arrayContaining([
-        "channels.discord.voice.tts.openai.apiKey",
-        "channels.discord.accounts.work.voice.tts.openai.apiKey",
+        "channels.discord.voice.tts.providers.openai.apiKey",
+        "channels.discord.accounts.work.voice.tts.providers.openai.apiKey",
       ]),
     );
   });
@@ -1974,8 +1976,10 @@ describe("secrets runtime snapshot", () => {
           discord: {
             voice: {
               tts: {
-                openai: {
-                  apiKey: { source: "env", provider: "default", id: "DISCORD_BASE_TTS_OPENAI" },
+                providers: {
+                  openai: {
+                    apiKey: { source: "env", provider: "default", id: "DISCORD_BASE_TTS_OPENAI" },
+                  },
                 },
               },
             },
@@ -1990,11 +1994,13 @@ describe("secrets runtime snapshot", () => {
                 enabled: true,
                 voice: {
                   tts: {
-                    openai: {
-                      apiKey: {
-                        source: "env",
-                        provider: "default",
-                        id: "DISCORD_ENABLED_OVERRIDE_TTS_OPENAI",
+                    providers: {
+                      openai: {
+                        apiKey: {
+                          source: "env",
+                          provider: "default",
+                          id: "DISCORD_ENABLED_OVERRIDE_TTS_OPENAI",
+                        },
                       },
                     },
                   },
@@ -2004,11 +2010,13 @@ describe("secrets runtime snapshot", () => {
                 enabled: false,
                 voice: {
                   tts: {
-                    openai: {
-                      apiKey: {
-                        source: "env",
-                        provider: "default",
-                        id: "DISCORD_DISABLED_OVERRIDE_TTS_OPENAI",
+                    providers: {
+                      openai: {
+                        apiKey: {
+                          source: "env",
+                          provider: "default",
+                          id: "DISCORD_DISABLED_OVERRIDE_TTS_OPENAI",
+                        },
                       },
                     },
                   },
@@ -2034,13 +2042,17 @@ describe("secrets runtime snapshot", () => {
       loadAuthStore: () => ({ version: 1, profiles: {} }),
     });
 
-    expect(snapshot.config.channels?.discord?.voice?.tts?.openai?.apiKey).toBe("base-tts-openai");
+    expect(snapshot.config.channels?.discord?.voice?.tts?.providers?.openai?.apiKey).toBe(
+      "base-tts-openai",
+    );
     expect(snapshot.config.channels?.discord?.pluralkit?.token).toBe("base-pk-token");
     expect(
-      snapshot.config.channels?.discord?.accounts?.enabledOverride?.voice?.tts?.openai?.apiKey,
+      snapshot.config.channels?.discord?.accounts?.enabledOverride?.voice?.tts?.providers?.openai
+        ?.apiKey,
     ).toBe("enabled-override-tts-openai");
     expect(
-      snapshot.config.channels?.discord?.accounts?.disabledOverride?.voice?.tts?.openai?.apiKey,
+      snapshot.config.channels?.discord?.accounts?.disabledOverride?.voice?.tts?.providers?.openai
+        ?.apiKey,
     ).toEqual({
       source: "env",
       provider: "default",
@@ -2055,7 +2067,7 @@ describe("secrets runtime snapshot", () => {
     );
     expect(snapshot.warnings.map((warning) => warning.path)).toEqual(
       expect.arrayContaining([
-        "channels.discord.accounts.disabledOverride.voice.tts.openai.apiKey",
+        "channels.discord.accounts.disabledOverride.voice.tts.providers.openai.apiKey",
         "channels.discord.accounts.disabledOverride.pluralkit.token",
       ]),
     );
@@ -2068,11 +2080,13 @@ describe("secrets runtime snapshot", () => {
           discord: {
             voice: {
               tts: {
-                openai: {
-                  apiKey: {
-                    source: "env",
-                    provider: "default",
-                    id: "DISCORD_UNUSED_BASE_TTS_OPENAI",
+                providers: {
+                  openai: {
+                    apiKey: {
+                      source: "env",
+                      provider: "default",
+                      id: "DISCORD_UNUSED_BASE_TTS_OPENAI",
+                    },
                   },
                 },
               },
@@ -2082,11 +2096,13 @@ describe("secrets runtime snapshot", () => {
                 enabled: true,
                 voice: {
                   tts: {
-                    openai: {
-                      apiKey: {
-                        source: "env",
-                        provider: "default",
-                        id: "DISCORD_ENABLED_ONLY_TTS_OPENAI",
+                    providers: {
+                      openai: {
+                        apiKey: {
+                          source: "env",
+                          provider: "default",
+                          id: "DISCORD_ENABLED_ONLY_TTS_OPENAI",
+                        },
                       },
                     },
                   },
@@ -2107,15 +2123,16 @@ describe("secrets runtime snapshot", () => {
     });
 
     expect(
-      snapshot.config.channels?.discord?.accounts?.enabledOverride?.voice?.tts?.openai?.apiKey,
+      snapshot.config.channels?.discord?.accounts?.enabledOverride?.voice?.tts?.providers?.openai
+        ?.apiKey,
     ).toBe("enabled-only-tts-openai");
-    expect(snapshot.config.channels?.discord?.voice?.tts?.openai?.apiKey).toEqual({
+    expect(snapshot.config.channels?.discord?.voice?.tts?.providers?.openai?.apiKey).toEqual({
       source: "env",
       provider: "default",
       id: "DISCORD_UNUSED_BASE_TTS_OPENAI",
     });
     expect(snapshot.warnings.map((warning) => warning.path)).toContain(
-      "channels.discord.voice.tts.openai.apiKey",
+      "channels.discord.voice.tts.providers.openai.apiKey",
     );
   });
 
@@ -2127,8 +2144,10 @@ describe("secrets runtime snapshot", () => {
             discord: {
               voice: {
                 tts: {
-                  openai: {
-                    apiKey: { source: "env", provider: "default", id: "DISCORD_BASE_TTS_OK" },
+                  providers: {
+                    openai: {
+                      apiKey: { source: "env", provider: "default", id: "DISCORD_BASE_TTS_OK" },
+                    },
                   },
                 },
               },
@@ -2137,11 +2156,13 @@ describe("secrets runtime snapshot", () => {
                   enabled: true,
                   voice: {
                     tts: {
-                      openai: {
-                        apiKey: {
-                          source: "env",
-                          provider: "default",
-                          id: "DISCORD_ENABLED_OVERRIDE_TTS_MISSING",
+                      providers: {
+                        openai: {
+                          apiKey: {
+                            source: "env",
+                            provider: "default",
+                            id: "DISCORD_ENABLED_OVERRIDE_TTS_MISSING",
+                          },
                         },
                       },
                     },
diff --git a/src/secrets/runtime.ts b/src/secrets/runtime.ts
index 5b29f36233b..38f71446b41 100644
--- a/src/secrets/runtime.ts
+++ b/src/secrets/runtime.ts
@@ -12,6 +12,7 @@ import {
   setRuntimeConfigSnapshot,
   type OpenClawConfig,
 } from "../config/config.js";
+import { migrateLegacyConfig } from "../config/legacy-migrate.js";
 import { resolveUserPath } from "../utils.js";
 import {
   collectCommandSecretAssignmentsFromSnapshot,
@@ -139,7 +140,9 @@ export async function prepareSecretsRuntimeSnapshot(params: {
 }): Promise<PreparedSecretsRuntimeSnapshot> {
   const runtimeEnv = mergeSecretsRuntimeEnv(params.env);
   const sourceConfig = structuredClone(params.config);
-  const resolvedConfig = structuredClone(params.config);
+  const resolvedConfig = structuredClone(
+    migrateLegacyConfig(params.config).config ?? params.config,
+  );
   const context = createResolverContext({
     sourceConfig,
     env: runtimeEnv,
diff --git a/src/tts/directives.ts b/src/tts/directives.ts
index df993ba1bed..ea1d7c76ef7 100644
--- a/src/tts/directives.ts
+++ b/src/tts/directives.ts
@@ -4,7 +4,6 @@ import { listSpeechProviders } from "./provider-registry.js";
 import type {
   SpeechModelOverridePolicy,
   SpeechProviderConfig,
-  SpeechProviderOverrides,
   TtsDirectiveOverrides,
   TtsDirectiveParseResult,
 } from "./provider-types.js";
@@ -13,7 +12,6 @@ type ParseTtsDirectiveOptions = {
   cfg?: OpenClawConfig;
   providers?: readonly SpeechProviderPlugin[];
   providerConfigs?: Record<string, SpeechProviderConfig>;
-  openaiBaseUrl?: string;
 };
 
 function buildProviderOrder(left: SpeechProviderPlugin, right: SpeechProviderPlugin): number {
@@ -36,49 +34,18 @@ function resolveDirectiveProviderConfig(
   provider: SpeechProviderPlugin,
   options?: ParseTtsDirectiveOptions,
 ): SpeechProviderConfig | undefined {
-  const explicit = options?.providerConfigs?.[provider.id];
-  if (explicit) {
-    return explicit;
-  }
-  if (provider.id === "openai" && options?.openaiBaseUrl) {
-    return { baseUrl: options.openaiBaseUrl };
-  }
-  return undefined;
-}
-
-function mergeProviderOverrides(
-  target: TtsDirectiveOverrides,
-  providerId: string,
-  next: SpeechProviderOverrides,
-): void {
-  target.providerOverrides = {
-    ...target.providerOverrides,
-    [providerId]: {
-      ...target.providerOverrides?.[providerId],
-      ...next,
-    },
-  };
-}
-
-function resolveLegacyOptions(
-  optionsOrOpenaiBaseUrl?: ParseTtsDirectiveOptions | string,
-): ParseTtsDirectiveOptions | undefined {
-  if (typeof optionsOrOpenaiBaseUrl === "string") {
-    return { openaiBaseUrl: optionsOrOpenaiBaseUrl };
-  }
-  return optionsOrOpenaiBaseUrl;
+  return options?.providerConfigs?.[provider.id];
 }
 
 export function parseTtsDirectives(
   text: string,
   policy: SpeechModelOverridePolicy,
-  optionsOrOpenaiBaseUrl?: ParseTtsDirectiveOptions | string,
+  options?: ParseTtsDirectiveOptions,
 ): TtsDirectiveParseResult {
   if (!policy.enabled) {
     return { cleanedText: text, overrides: {}, warnings: [], hasDirective: false };
   }
 
-  const options = resolveLegacyOptions(optionsOrOpenaiBaseUrl);
   const providers = resolveDirectiveProviders(options);
   const overrides: TtsDirectiveOverrides = {};
   const warnings: string[] = [];
@@ -135,7 +102,13 @@ export function parseTtsDirectives(
         }
         handled = true;
         if (parsed.overrides) {
-          mergeProviderOverrides(overrides, provider.id, parsed.overrides);
+          overrides.providerOverrides = {
+            ...overrides.providerOverrides,
+            [provider.id]: {
+              ...overrides.providerOverrides?.[provider.id],
+              ...parsed.overrides,
+            },
+          };
         }
         if (parsed.warnings?.length) {
           warnings.push(...parsed.warnings);
diff --git a/src/tts/tts.test.ts b/src/tts/tts.test.ts
index d78e9494bca..a3ae013a779 100644
--- a/src/tts/tts.test.ts
+++ b/src/tts/tts.test.ts
@@ -86,7 +86,6 @@ const {
   parseTtsDirectives,
   resolveModelOverridePolicy,
   summarizeText,
-  resolveOutputFormat,
   getResolvedSpeechProviderConfig,
 } = _test;
 
@@ -234,65 +233,6 @@ describe("tts", () => {
     });
   });
 
-  describe("resolveOutputFormat", () => {
-    it("selects opus for opus channels (telegram/feishu/whatsapp/matrix) and mp3 for others", () => {
-      const cases = [
-        {
-          channel: "telegram",
-          expected: {
-            openai: "opus",
-            elevenlabs: "opus_48000_64",
-            extension: ".opus",
-            voiceCompatible: true,
-          },
-        },
-        {
-          channel: "feishu",
-          expected: {
-            openai: "opus",
-            elevenlabs: "opus_48000_64",
-            extension: ".opus",
-            voiceCompatible: true,
-          },
-        },
-        {
-          channel: "whatsapp",
-          expected: {
-            openai: "opus",
-            elevenlabs: "opus_48000_64",
-            extension: ".opus",
-            voiceCompatible: true,
-          },
-        },
-        {
-          channel: "matrix",
-          expected: {
-            openai: "opus",
-            elevenlabs: "opus_48000_64",
-            extension: ".opus",
-            voiceCompatible: true,
-          },
-        },
-        {
-          channel: "discord",
-          expected: {
-            openai: "mp3",
-            elevenlabs: "mp3_44100_128",
-            extension: ".mp3",
-            voiceCompatible: false,
-          },
-        },
-      ] as const;
-      for (const testCase of cases) {
-        const output = resolveOutputFormat(testCase.channel);
-        expect(output.openai, testCase.channel).toBe(testCase.expected.openai);
-        expect(output.elevenlabs, testCase.channel).toBe(testCase.expected.elevenlabs);
-        expect(output.extension, testCase.channel).toBe(testCase.expected.extension);
-        expect(output.voiceCompatible, testCase.channel).toBe(testCase.expected.voiceCompatible);
-      }
-    });
-  });
-
   describe("resolveEdgeOutputFormat", () => {
     const baseCfg: OpenClawConfig = {
       agents: { defaults: { model: { primary: "openai/gpt-4o-mini" } } },
@@ -383,9 +323,11 @@ describe("tts", () => {
     it("accepts custom voices and models when openaiBaseUrl is a non-default endpoint", () => {
       const policy = resolveModelOverridePolicy({ enabled: true });
       const input = "Hello [[tts:voice=kokoro-chinese model=kokoro-v1]] world";
-      const customBaseUrl = "http://localhost:8880/v1";
-
-      const result = parseTtsDirectives(input, policy, customBaseUrl);
+      const result = parseTtsDirectives(input, policy, {
+        providerConfigs: {
+          openai: { baseUrl: "http://localhost:8880/v1" },
+        },
+      });
       const openaiOverrides = result.overrides.providerOverrides?.openai as
         | { voice?: string; model?: string }
         | undefined;
@@ -398,9 +340,11 @@ describe("tts", () => {
     it("rejects unknown voices and models when openaiBaseUrl is the default OpenAI endpoint", () => {
       const policy = resolveModelOverridePolicy({ enabled: true });
       const input = "Hello [[tts:voice=kokoro-chinese model=kokoro-v1]] world";
-      const defaultBaseUrl = "https://api.openai.com/v1";
-
-      const result = parseTtsDirectives(input, policy, defaultBaseUrl);
+      const result = parseTtsDirectives(input, policy, {
+        providerConfigs: {
+          openai: { baseUrl: "https://api.openai.com/v1" },
+        },
+      });
       const openaiOverrides = result.overrides.providerOverrides?.openai as
         | { voice?: string }
         | undefined;
diff --git a/src/tts/tts.ts b/src/tts/tts.ts
index 479c2ecd129..f3f32b438db 100644
--- a/src/tts/tts.ts
+++ b/src/tts/tts.ts
@@ -48,22 +48,6 @@ const DEFAULT_TTS_MAX_LENGTH = 1500;
 const DEFAULT_TTS_SUMMARIZE = true;
 const DEFAULT_MAX_TEXT_LENGTH = 4096;
 
-const OPUS_OUTPUT = {
-  openai: "opus" as const,
-  // ElevenLabs output formats use codec_sample_rate_bitrate naming.
-  // Opus @ 48kHz/64kbps is a good voice message tradeoff.
-  elevenlabs: "opus_48000_64",
-  extension: ".opus",
-  voiceCompatible: true,
-};
-
-const DEFAULT_OUTPUT = {
-  openai: "mp3" as const,
-  elevenlabs: "mp3_44100_128",
-  extension: ".mp3",
-  voiceCompatible: false,
-};
-
 export type ResolvedTtsConfig = {
   auto: TtsAutoMode;
   mode: TtsMode;
@@ -418,16 +402,9 @@ export function setLastTtsAttempt(entry: TtsStatusEntry | undefined): void {
   lastTtsAttempt = entry;
 }
 
-/** Channels that require opus audio */
+/** Channels that require voice-note-compatible audio */
 const OPUS_CHANNELS = new Set(["telegram", "feishu", "whatsapp", "matrix"]);
 
-function resolveOutputFormat(channelId?: string | null) {
-  if (channelId && OPUS_CHANNELS.has(channelId)) {
-    return OPUS_OUTPUT;
-  }
-  return DEFAULT_OUTPUT;
-}
-
 function resolveChannelId(channel: string | undefined): ChannelId | null {
   return channel ? normalizeChannelId(channel) : null;
 }
@@ -876,6 +853,5 @@ export const _test = {
   parseTtsDirectives,
   resolveModelOverridePolicy,
   summarizeText,
-  resolveOutputFormat,
   getResolvedSpeechProviderConfig,
 };