chenhg5/cc-connect · error

send audio: %w

Error message

send audio: %w

What it means

cc-connect wraps failures from the platform adapter's SendAudio call (delivering synthesized TTS audio to the messaging platform) with the 'send audio: %w' prefix. The TTS synthesis itself succeeded; the failure occurs while transmitting the audio bytes to the platform (e.g. Feishu, Telegram). The platform-specific cause is preserved via %w.

Source

Thrown at core/engine.go:15893

		return fmt.Errorf("text exceeds max_text_len (%d > %d)", utf8.RuneCountInString(text), e.tts.MaxTextLen)
	}
	as, ok := p.(AudioSender)
	if !ok {
		return fmt.Errorf("platform %s does not support audio sending", p.Name())
	}
	slog.Info("tts: starting synthesis", "voice", e.tts.Voice, "speed", e.tts.Speed, "text_len", len(text))
	opts := TTSSynthesisOpts{
		Voice:        e.tts.Voice,
		LanguageType: e.tts.LanguageType,
		Speed:        e.tts.Speed,
	}
	audioData, format, err := e.tts.TTS.Synthesize(e.ctx, StripMarkdown(text), opts)
	if err != nil {
		return fmt.Errorf("synthesize: %w", err)
	}
	slog.Info("tts: synthesis successful", "format", format, "audio_size", len(audioData))
	if err := as.SendAudio(e.ctx, replyCtx, audioData, format); err != nil {
		return fmt.Errorf("send audio: %w", err)
	}
	slog.Info("tts: audio sent successfully", "platform", p.Name())
	return nil
}

// ──────────────────────────────────────────────────────────────
// Bot-to-bot relay
// ──────────────────────────────────────────────────────────────

type platformNameOnly struct {
	name string
}

func (p platformNameOnly) Name() string                           { return p.name }
func (platformNameOnly) Start(MessageHandler) error               { return nil }
func (platformNameOnly) Reply(context.Context, any, string) error { return nil }
func (platformNameOnly) Send(context.Context, any, string) error  { return nil }
func (platformNameOnly) Stop() error                              { return nil }

View on GitHub (pinned to 4000b2338a)

Solutions

1. Inspect the wrapped cause for the platform-specific failure (auth vs size vs format)
2. Re-authenticate or refresh the platform bot token in config.toml
3. Check the platform's max audio/file size and the produced audio_size in the preceding log line
4. Convert the TTS output to a format the platform accepts (e.g. opus/ogg for Telegram voice)
5. Retry — upload failures are often transient

Example fix

// before
if err := as.SendAudio(e.ctx, replyCtx, audioData, format); err != nil {
    return fmt.Errorf("send audio: %w", err)
}
// after
if err := as.SendAudio(e.ctx, replyCtx, audioData, format); err != nil {
    slog.Warn("tts: audio send failed, falling back to text", "platform", p.Name(), "error", err)
    if ferr := e.replyText(replyCtx, text); ferr != nil {
        return fmt.Errorf("send audio: %w (text fallback also failed: %v)", err, ferr)
    }
    return nil
}

// check platform capability before attempting audio
if _, ok := as.(interface{ SendAudio(ctx context.Context, rc ReplyContext, data []byte, format string) error }); !ok {
    return e.replyText(replyCtx, text)
}

if err := as.SendAudio(e.ctx, replyCtx, audioData, format); err != nil {
    slog.Warn("send audio failed; falling back to text", "platform", p.Name(), "error", err)
    return e.replyText(replyCtx, text)
}

When it happens

Trigger: After successful Synthesize, as.SendAudio(e.ctx, replyCtx, audioData, format) returns an error — the platform API rejects the upload (auth token expired, file too large, unsupported audio format for that platform, network failure, or the chat/message target is invalid).

Common situations: Platform bot token expired or rotated; audio file exceeds the platform's upload size limit; audio format from TTS (e.g. mp3) not accepted by the platform; chat ID no longer valid (user blocked bot); transient network failure during upload.

Understand the failure class

Background: "API request failed": what wrapped HTTP errors from external APIs mean and how to find the real cause — this error's family across 29 libraries.

Related errors

• synthesize: %w
• openai tts: create request: %w
• openai tts: request: %w
• openai tts API %d: %s
• openai tts: read audio: %w

AI-assisted analysis of chenhg5/cc-connect@4000b2338a (2026-09-06). Data as JSON: /api/errors/d7e60595a6b624e4. Report an issue: GitHub.