PR #4: fix(speech-core): add discord to OPUS_CHANNELS for native voice-note TTS

• Repository: EronFan/openclaw-fork
• Author: EronFan
• State: open | merged: False
• Link: https://github.com/EronFan/openclaw-fork/pull/4

Description (problem / solution / changelog)

Summary

Fixes #63931 — Discord missing from OPUS_CHANNELS breaks native auto voice replies.

Root Cause

In extensions/speech-core/src/tts.ts, the OPUS_CHANNELS Set did not include "discord". This caused TTS for Discord to fall back to MP3 attachment instead of using the native voice-note path (Opus codec).

const OPUS_CHANNELS = new Set(["telegram", "feishu", "whatsapp", "matrix", "discord"]);

(Added "discord" to the existing set of channels.)

Fix

Added "discord" to the OPUS_CHANNELS Set. The author of #63931 has locally verified this fix works.

Testing

The author confirmed the fix locally. CI should validate the change.

Closes #63931

Changed files

• extensions/speech-core/src/tts.ts (modified, +1/-1)
• extensions/telegram/src/outbound-adapter.test.ts (modified, +82/-0)
• extensions/telegram/src/outbound-adapter.ts (modified, +14/-0)
• src/agents/pi-tools.ts (modified, +5/-1)
• src/agents/tool-policy.test.ts (modified, +24/-0)
• src/agents/tool-policy.ts (modified, +19/-2)

PR #63950: fix(speech-core): add discord to OPUS_CHANNELS for native voice-note TTS

• Repository: openclaw/openclaw
• Author: EronFan
• State: closed | merged: False
• Link: https://github.com/openclaw/openclaw/pull/63950

Description (problem / solution / changelog)

Summary

Fixes #63931 — Discord missing from OPUS_CHANNELS breaks native auto voice replies.

Root Cause

In extensions/speech-core/src/tts.ts, the OPUS_CHANNELS Set did not include "discord". This caused TTS for Discord to fall back to MP3 attachment instead of using the native voice-note path (Opus codec).

const OPUS_CHANNELS = new Set(["telegram", "feishu", "whatsapp", "matrix", "discord"]);

(Added "discord" to the existing set of channels.)

Fix

Added "discord" to the OPUS_CHANNELS Set. The author of #63931 has locally verified this fix works.

Testing

The author confirmed the fix locally. CI should validate the change.

Closes #63931

Changed files

• extensions/speech-core/src/tts.ts (modified, +1/-1)
• src/agents/pi-tools.ts (modified, +5/-1)
• src/agents/tool-policy.test.ts (modified, +24/-0)
• src/agents/tool-policy.ts (modified, +19/-2)

PR #64045: fix: add discord to OPUS_CHANNELS to enable native auto voice replies

• Repository: openclaw/openclaw
• Author: EronFan
• State: closed | merged: False
• Link: https://github.com/openclaw/openclaw/pull/64045

Description (problem / solution / changelog)

Summary

Add discord to the OPUS_CHANNELS set in extensions/speech-core/src/tts.ts so that Discord can use native Opus-encoded voice replies instead of falling back to audio-file attachments.

Fix

-const OPUS_CHANNELS = new Set(["telegram", "feishu", "whatsapp", "matrix"]);
+const OPUS_CHANNELS = new Set(["telegram", "feishu", "whatsapp", "matrix", "discord"]);

Issue

Fixes #63931

Testing

Author locally verified.

Changed files

• extensions/discord/src/shared.ts (modified, +2/-1)
• extensions/speech-core/src/tts.ts (modified, +1/-1)
• src/agents/pi-tools.ts (modified, +5/-1)
• src/agents/tool-policy.test.ts (modified, +24/-0)
• src/agents/tool-policy.ts (modified, +19/-2)
• src/media-understanding/openai-compatible-audio.ts (modified, +3/-0)

PR #64096: fix(speech-core): add discord to OPUS_CHANNELS for native voice-note TTS

• Repository: openclaw/openclaw
• Author: LiuHuaize
• State: closed | merged: True
• Link: https://github.com/openclaw/openclaw/pull/64096

Description (problem / solution / changelog)

Summary

• Problem: extensions/speech-core/src/tts.ts did not treat Discord as an Opus-capable voice-note channel in the auto-TTS path.
• Why it matters: Discord auto voice replies fell back to regular audio attachments instead of native voice messages.
• What changed: added discord to OPUS_CHANNELS and added a focused regression test that locks in Discord auto-TTS voice-note routing.
• What did NOT change (scope boundary): no Discord delivery code, provider-specific synthesis logic, or non-Discord channel behavior changed.

Change Type (select all)

• Bug fix
• Feature
• Refactor required for the fix
• Docs
• Security hardening
• Chore/infra

Scope (select all touched areas)

• Gateway / orchestration
• Skills / tool execution
• Auth / tokens
• Memory / storage
• Integrations
• API / contracts
• UI / DX
• CI/CD / infra

Linked Issue/PR

• Closes #63931
• Related #63950
• This PR fixes a bug or regression

Root Cause (if applicable)

• Root cause: the speech-core runtime chose target: "audio-file" for Discord because discord was missing from OPUS_CHANNELS.
• Missing detection / guardrail: there was no regression test asserting that Discord auto-TTS uses the voice-note path.
• Contributing context (if known): #63950 already had the same one-line runtime fix, but that PR was closed for PR-queue reasons rather than because the fix was incorrect.

Regression Test Plan (if applicable)

• Coverage level that should have caught this:
  • Unit test
  • Seam / integration test
  • End-to-end test
  • Existing coverage already sufficient
• Target test or file: extensions/speech-core/src/tts.test.ts
• Scenario the test should lock in: Discord auto TTS should synthesize with target: "voice-note" and return audioAsVoice: true for a voice-compatible result.
• Why this is the smallest reliable guardrail: it exercises the exact speech-core decision point that controls Discord's auto voice-note routing without needing live Discord or provider credentials.
• Existing test that already covers this (if any): None.
• If no new test is added, why not: N/A

User-visible / Behavior Changes

Discord auto TTS / auto voice replies now route through the native voice-note path instead of falling back to regular audio attachments when the synthesis result is voice-compatible.

Diagram (if applicable)

Before:
[Discord auto TTS] -> [channel not in OPUS_CHANNELS] -> [target: audio-file] -> [regular audio attachment]

After:
[Discord auto TTS] -> [channel in OPUS_CHANNELS] -> [target: voice-note] -> [native Discord voice message]

Security Impact (required)

• New permissions/capabilities? (No)
• Secrets/tokens handling changed? (No)
• New/changed network calls? (No)
• Command/tool execution surface changed? (No)
• Data access scope changed? (No)
• If any Yes, explain risk + mitigation:

Repro + Verification

Environment

• OS: macOS (darwin arm64)
• Runtime/container: Node 22 / pnpm workspace
• Model/provider: mocked speech provider in unit test
• Integration/channel (if any): Discord
• Relevant config (redacted): messages.tts.enabled=true, messages.tts.provider=mock

Steps

1. Run pnpm test extensions/speech-core/src/tts.test.ts.
2. Let the test call maybeApplyTtsToPayload for channel: "discord" with a mock provider that only marks voice output as compatible when target === "voice-note".
3. Assert the synthesize request uses target: "voice-note" and the resulting payload sets audioAsVoice: true.

Expected

• Discord auto TTS uses the voice-note path.
• The returned payload is marked for native voice delivery.

Actual

• Verified locally after the fix.

Evidence

Attach at least one:

• Failing test/log before + passing after
• Trace/log snippets
• Screenshot/recording
• Perf numbers (if relevant)

Human Verification (required)

• Verified scenarios: added and ran pnpm test extensions/speech-core/src/tts.test.ts locally; the test passed and asserted both target: "voice-note" and audioAsVoice: true for Discord.
• Edge cases checked: the mock provider only reports voiceCompatible when the request target is the voice-note path, so the test guards both the target selection and the final payload flag.
• What you did not verify: live Discord delivery, full repo-wide pnpm check, and broader integration suites.

Review Conversations

• I replied to or resolved every bot review conversation I addressed in this PR.
• I left unresolved only the conversations that still need reviewer or maintainer judgment.

Compatibility / Migration

• Backward compatible? (Yes)
• Config/env changes? (No)
• Migration needed? (No)
• If yes, exact upgrade steps:

Risks and Mitigations

• Risk: the regression test mocks channel normalization instead of booting the full channel registry.
  • Mitigation: the test intentionally isolates the speech-core decision path and asserts the exact voice-note target and payload flag that regressed here.

Changed files

• CHANGELOG.md (modified, +1/-0)
• extensions/speech-core/src/tts.test.ts (added, +99/-0)
• extensions/speech-core/src/tts.ts (modified, +1/-1)