fix(telegram): combine voice payload with preceding streamed text by sdimantsd · Pull Request #81131 · openclaw/openclaw

sdimantsd · 2026-05-12T19:27:19Z

Summary

When a final voice payload arrives in Telegram after partial/block streaming has already sent the text, two separate Telegram messages appear: first the streamed text, then the voice note. This restores the pre-bca16d0f00 behavior where audioAsVoice payloads implicitly carried the spoken text as the voice's caption, so voice+text arrive as one combined message.

Behavior change

Before: User receives 1) streamed text message 2) separate voice note (two messages).
After: User receives a single voice note with the spoken text as caption.

Real behavior proof

After-fix screenshot (Telegram chat, agent reply with TTS):

The single message in the screenshot is the voice note with the spoken text bound to it as caption.

Running on openclaw:local built from this branch on top of 2026.5.12-beta.1. Patch identifiable at runtime in /app/dist/bot-DiUeqsFG.js (grep -c voice-combine returns 2).

Mechanism

When info.kind === final, effectivePayload.audioAsVoice === true, and answerLane.hasStreamedMessage:

Delete the previously-streamed text message via bot.api.deleteMessage.
Rebuild payloadForSend with text set to the spoken text — the bot send pipeline then attaches it as caption to the voice.
Clear answerLane.hasStreamedMessage and the partial-text trackers.

The keying on audioAsVoice + hasStreamedMessage is necessary because audioAsVoice payloads don't set mediaUrls (so reply.hasMedia=false), and in partial-streaming mode the payload carries text (so reply.trimmedText is non-empty).

Test plan

Local reproduction confirmed: streaming text + voice payload arrives as a single voice message with caption (see screenshot).
Maintainer review of keying logic appreciated — happy to widen/narrow the condition if there are edge cases.

When a final voice payload arrives after partial/block streaming already sent the text, delete the streamed text message and attach the spoken text as caption so voice+text arrive as one combined message. This restores pre-bca16d0f00 behavior where audioAsVoice payloads implicitly carried the spoken text and the streamed text was suppressed. Note: audioAsVoice payloads don't set mediaUrls (reply.hasMedia=false) and in partial streaming mode the payload carries text (reply.trimmedText non-empty), so we key only on audioAsVoice + hasStreamedMessage.

clawsweeper · 2026-05-12T19:30:34Z

Thanks for the context here. I swept through the related work, and this is now duplicate or superseded.

Close this PR as superseded: the remaining Telegram final-mode TTS behavior is now tracked by the canonical issue and broader replacement PR, while this one-file Telegram transport patch is the riskier delete-and-resend path.

Canonical path: Close this branch and continue the captioned final-mode TTS decision and implementation review in #83511 and #83988.

So I’m closing this here and keeping the remaining discussion on #83511 and #83988.

Review details

Best possible solution:

Close this branch and continue the captioned final-mode TTS decision and implementation review in #83511 and #83988.

Do we have a high-confidence way to reproduce the issue?

Yes. Source and tests on current main show late media after streamed Telegram text is delivered as separate media while the streamed text remains, and the PR screenshot shows the proposed combined voice-caption result; I did not run a fresh live Telegram session in this read-only review.

Is this the best way to solve the issue?

No. This is not the best fix because it deletes an already-delivered Telegram message in the transport layer; the safer canonical path is the upstream capability-gated final-TTS deferral under review.

Security review:

Security review cleared: No concrete security or supply-chain concern was found; the diff only changes Telegram runtime delivery behavior.

AGENTS.md: found and applied where relevant.

What I checked:

Repository policy applied: Root AGENTS.md was read fully from disk; the review used its guidance on full-path PR review, Telegram/channel proof, and compatibility-sensitive message delivery changes. (AGENTS.md:1, fef839407995)
Scoped extension policy applied: The extension boundary guide was read and applied because the touched file is under a bundled plugin. (extensions/AGENTS.md:1, fef839407995)
Telegram maintainer note applied: The Telegram review note requires real Telegram proof for transport/streaming behavior and treats a single visible preview message as intentional streaming behavior. (.agents/maintainer-notes/telegram.md:1, fef839407995)
Current main owns late media behavior elsewhere: Current main routes late media after streamed text through createLaneTextDeliverer, finalizing the preview and sending a media-only payload rather than deleting the visible text in the dispatch fallback branch. (extensions/telegram/src/lane-delivery-text-deliverer.ts:544, fef839407995)
Current main tests the separate late-media behavior: The Telegram dispatch test explicitly expects the streamed text to remain and the late media payload to be delivered without text. (extensions/telegram/src/bot-message-dispatch.test.ts:2582, fef839407995)
Current main does not treat spokenText as a successful caption: The delivery test asserts spokenText is not rendered as a caption on a successful voice send, so changing that behavior is a delivery-policy decision rather than a small transport cleanup. (extensions/telegram/src/bot/delivery.test.ts:1269, fef839407995)

Likely related people:

obviyus: GitHub commit metadata and blame connect this person to the current lane-delivery behavior and the earlier streamed-text finalization change that this PR tries to alter. (role: introduced/current behavior contributor; confidence: high; commits: d498b1cce49c, bca16d0f00c0; files: extensions/telegram/src/lane-delivery-text-deliverer.ts, extensions/telegram/src/bot-message-dispatch.ts, extensions/telegram/src/bot-message-dispatch.test.ts)
ragesaq: Recent current-main history touches nearby auto-reply/channel final delivery behavior that intersects with the final-mode TTS delivery path. (role: recent adjacent contributor; confidence: medium; commits: a15427d605fe; files: src/auto-reply/reply/dispatch-acp.ts, src/auto-reply/reply/dispatch-from-config.ts, extensions/telegram/src/shared.ts)
Jerry-Xin: Authored the open broader PR that currently owns the captioned final-mode TTS implementation path for the canonical issue. (role: active replacement PR author; confidence: high; commits: e2de93d4f6ae; files: src/auto-reply/reply/dispatch-acp.ts, src/auto-reply/reply/dispatch-from-config.ts, extensions/telegram/src/shared.ts)

Codex review notes: model internal, reasoning high; reviewed against fef839407995.

…se on the fork)

openclaw-barnacle · 2026-05-28T04:48:18Z

This pull request has been automatically marked as stale due to inactivity.
Please add updates or it will be closed.

clawsweeper · 2026-05-28T04:56:41Z

ClawSweeper PR egg: 🔥 warming; proof passed, review follow-up or readiness checks remain. Hatch with @clawsweeper hatch when eligible.

Rules and details

Hatchability:

Merged PRs are hatchable.
Open PRs are hatchable when they are status: 👀 ready for maintainer look, status: 🚀 automerge armed, or labeled clawsweeper:automerge.
Closed unmerged PRs are hatchable only when one of those hatchable labels is still present in the durable record.

About:

Eggs appear after real-behavior proof passes. They are collectible flavor only.
Review momentum changes the shell state: follow-up work warms it, re-review makes it wobble, and a clean final review lets it hatch.
The hatch is seeded from this repository and PR number, so the same PR keeps the same creature; the reviewed head SHA can only change safe visual details.
Rarity is just collectible sparkle: 🥚 common, 🌱 uncommon, 💎 rare, ✨ glimmer, and 🌈 legendary.

openclaw-clownfish · 2026-06-15T23:05:40Z

Thanks @sdimantsd for the focused Telegram proof and patch in #81131. I am closing this as superseded by #83988 because the remaining final-mode TTS churn is now being handled in the broader canonical path for #83511.

Clownfish is keeping the canonical PR open there so implementation review, validation, and follow-up stay in one place. The source PR context and attribution are preserved in the cluster record, and your Telegram repro/proof remains useful for the canonical review. If this branch covers a different reproduction path or the issue still reproduces after #83988 lands, please reply and we can reopen or split it back out.

openclaw-barnacle Bot added channel: telegram Channel integration: telegram size: XS triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. labels May 12, 2026

clawsweeper Bot added the mantis: telegram-visible-proof Mantis should capture Telegram visible proof. label May 12, 2026

Add after-fix screenshot for PR openclaw#81131

6de76e7

clawsweeper Bot added the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 13, 2026

openclaw-barnacle Bot removed the proof: sufficient ClawSweeper judged the real behavior proof convincing. label May 13, 2026

chore: move proof screenshot out of source diff (referenced via relea…

df74b4e

…se on the fork)

clawsweeper Bot mentioned this pull request May 21, 2026

Feature: TTS delivery mode option — separate text + voice messages on Telegram #42539

Open

openclaw-barnacle Bot added the stale Marked as stale due to inactivity label May 28, 2026

openclaw-barnacle Bot removed triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. stale Marked as stale due to inactivity labels May 28, 2026

openclaw-barnacle Bot added the triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. label May 29, 2026

clawsweeper Bot had a problem deploying to qa-live-shared May 29, 2026 05:03 Failure

openclaw-barnacle Bot removed the triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. label May 29, 2026

openclaw-barnacle Bot added the triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. label May 29, 2026

openclaw-barnacle Bot removed the triage: needs-real-behavior-proof Candidate: external PR needs after-fix proof from a real setup. label Jun 14, 2026

clawsweeper Bot had a problem deploying to qa-live-shared June 14, 2026 21:54 Failure

openclaw-clownfish Bot closed this Jun 15, 2026

openclaw-clownfish Bot added the clownfish Tracked by Clownfish automation label Jun 15, 2026

This was referenced Jun 16, 2026

fix(tts): defer text settlement for final-mode TTS to eliminate churn (#83511) #83988

Open

TTS in final mode: visible text sent before audio causes Telegram delete-and-resend churn #83511

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

fix(telegram): combine voice payload with preceding streamed text#81131

fix(telegram): combine voice payload with preceding streamed text#81131
sdimantsd wants to merge 3 commits into
openclaw:mainfrom
sdimantsd:fix/telegram-voice-caption-combined

sdimantsd commented May 12, 2026 •

edited

Loading

Uh oh!

clawsweeper Bot commented May 12, 2026 •

edited

Loading

Uh oh!

openclaw-barnacle Bot commented May 28, 2026

Uh oh!

clawsweeper Bot commented May 28, 2026

Uh oh!

openclaw-clownfish Bot commented Jun 15, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Uh oh!

Conversation

sdimantsd commented May 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Behavior change

Real behavior proof

Mechanism

Test plan

Uh oh!

clawsweeper Bot commented May 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

openclaw-barnacle Bot commented May 28, 2026

Uh oh!

clawsweeper Bot commented May 28, 2026

Uh oh!

openclaw-clownfish Bot commented Jun 15, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

sdimantsd commented May 12, 2026 •

edited

Loading

clawsweeper Bot commented May 12, 2026 •

edited

Loading