Transcribe audio files to text via the Z.AI GLM-ASR API — POST /audio/transcriptions with a wav/mp3 (≤25MB, ≤30s). Triggers: transcribe, speech to text, ASR, audio transcription.
Scanned 9/20/2026
Install to Claude Code
npx -y skills add darellchua2/opencode-config-template --skill zai-asr-skill --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Zai Asr Skill?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/darellchua2-zai-asr-skill)More formats (shields.io, HTML) on the badges page.
---
name: zai-asr-skill
description: >-
Transcribe audio files to text via the Z.AI GLM-ASR API — POST
/audio/transcriptions with a wav/mp3 (≤25MB, ≤30s). Triggers: transcribe,
speech to text, ASR, audio transcription.
license: Apache-2.0
compatibility: opencode
category: Media Generation
---
## What I do
I provide the exact recipe for an agent to **transcribe an audio file** via the Z.AI
**GLM-ASR** API and return the **transcript text**. OpenCode has no voice-input path,
so transcription happens as a file-based tool call: `POST /audio/transcriptions`.
## Constraints (hard limits — check before calling)
- Models: `glm-asr-2512`
- File: **≤ 25 MB**, **≤ 30 seconds**, `.wav` or `.mp3`
- Longer audio must be split first (e.g. `ffmpeg -f segment -segment_times <n>`) —
transcribe the segments in order and concatenate.
## Prerequisite — API key resolution
```bash
KEY="${ZAI_API_KEY:-$(jq -r '.["zai"].key // .["zai-coding-plan"].key // empty' \
~/.local/share/opencode/auth.json 2>/dev/null)}"
[ -z "$KEY" ] && { echo "ZAI_API_KEY not found — set it (export ZAI_API_KEY) or run \`opencode auth login\` (Z.AI)."; exit 1; }
```
If neither source has the key, **stop and report** — do not proceed.
## Recipe
```bash
# --- options ---
FILE="path/to/audio.mp3" # REQUIRED — .wav or .mp3, ≤25MB, ≤30s
BASE="${ZAI_MEDIA_ENDPOINT:-https://api.z.ai/api/paas/v4}" # pay-as-you-go ONLY
# pre-flight: size + duration guards
[ -f "$FILE" ] || { echo "No such file: $FILE"; exit 1; }
[ "$(wc -c <"$FILE")" -le 26214400 ] || { echo "File >25MB — split it first"; exit 1; }
RESP=$(curl -sS --max-time 120 -X POST "$BASE/audio/transcriptions" \
-H "Authorization: Bearer $KEY" \
-F "model=glm-asr-2512" \
-F "file=@$FILE")
TEXT=$(printf '%s' "$RESP" | python3 -c '
import sys, json
d = json.load(sys.stdin)
if d.get("error"):
sys.stderr.write("API error: " + json.dumps(d["error"]) + "\n"); sys.exit(1)
print(d.get("text",""))') || { echo "Transcription failed. Response: $RESP"; exit 1; }
echo "TRANSCRIPT: $TEXT"
```
**Output:** the transcript on the `TRANSCRIPT:` line.
## Error handling
| Condition | Detect | Action |
|-----------|--------|--------|
| Missing key | `[ -z "$KEY" ]` | Stop; tell caller to set `ZAI_API_KEY` |
| File too large / long | pre-flight check | Split with ffmpeg; transcribe segments; concatenate |
| Wrong format | file extension not wav/mp3 | Convert: `ffmpeg -i in -ar 16000 out.wav` |
| API error (auth/quota/format) | response JSON has `"error"` | Print error body; report `Status: failed` |
| Empty transcript | `text` empty | File may be silent — report, don't retry |
| Rate limited | HTTP 429 | Wait and retry once; otherwise report |
## Cost note
GLM-ASR is **pay-as-you-go** (~$0.03/MTok), **not** covered by the GLM Coding Plan —
the recipe targets `/api/paas/v4`.
## Caller contract
After running the recipe, return the **transcript text** (the `TRANSCRIPT:` line).
Do not echo API response JSON. If transcription failed, return `Status: failed`
with the error message and stop.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!