Video to Text Transcription — Convert Video Speech to Text

## Getting Started

> Send me your video files and I'll handle the AI speech transcription. Or just describe what you're after.

**Try saying:**
- "convert a 10-minute interview recorded on a smartphone into a 1080p MP4"
- "transcribe the spoken dialogue into a text document"
- "converting spoken video content into editable text documents for journalists, students, content creators"

### First-Time Connection

When a user first opens this skill, connect to the processing backend automatically. Briefly let them know (e.g. "Setting up...").

**Authentication**: Check if `NEMO_TOKEN` is set in the environment. If it is, skip to step 2.

1. **Obtain a free token**: Generate a random UUID as client identifier. POST to `https://mega-api-prod.nemovideo.ai/api/auth/anonymous-token` with header `X-Client-Id` set to that UUID. The response `data.token` is your NEMO_TOKEN — 100 free credits, valid 7 days.
2. **Create a session**: POST to `https://mega-api-prod.nemovideo.ai/api/tasks/me/with-session/nemo_agent` with `Authorization: Bearer <token>`, `Content-Type: application/json`, and body `{"task_name":"project","language":"<detected>"}`. Store the returned `session_id` for all subsequent requests.

Keep setup communication brief. Don't display raw API responses or token values to the user.

# Video to Text Transcription — Convert Video Speech to Text

This tool takes your video files and runs AI speech transcription through a cloud rendering pipeline. You upload, describe what you want, and download the result.

Say you have a 10-minute interview recorded on a smartphone and want to transcribe the spoken dialogue into a text document — the backend processes it in about 1-2 minutes and hands you a 1080p MP4.

Tip: shorter clips under 5 minutes produce faster and more accurate transcripts.

## Matching Input to Actions

User prompts referencing video to text transcription, aspect ratio, text overlays, or audio tracks get routed to the corresponding action via keyword and intent classification.

| User says... | Action | Skip SSE? |
|-------------|--------|----------|
| "export" / "导出" / "download" / "send me the video" | → §3.5 Export | ✅ |
| "credits" / "积分" / "balance" / "余额" | → §3.3 Credits | ✅ |
| "status" / "状态" / "show tracks" | → §3.4 State | ✅ |
| "upload" / "上传" / user sends file | → §3.2 Upload | ✅ |
| Everything else (generate, edit, add BGM…) | → §3.1 SSE | ❌ |

## Cloud Render Pipeline Details

Each export job queues on a cloud GPU node that composites video layers, applies platform-spec compression (H.264, up to 1080x1920), and returns a download URL within 30-90 seconds. The session token carries render job IDs, so closing the tab before completion orphans the job.

Every API call needs `Authorization: Bearer <NEMO_TOKEN>` plus the three attribution headers above. If any header is missing, exports return 402.

Headers are derived from this file's YAML frontmatter. `X-Skill-Source` is `video-to-text-transcription`, `X-Skill-Version` comes from the `version` field, and `X-Skill-Platform` is detected from the install path (`~/.clawhub/` = `clawhub`, `~/.cursor/skills/` = `cursor`, otherwise `unknown`).

**API base**: `https://mega-api-prod.nemovideo.ai`

**Create session**: POST `/api/tasks/me/with-session/nemo_agent` — body `{"task_name":"project","language":"<lang>"}` — returns `task_id`, `session_id`.

**Send message (SSE)**: POST `/run_sse` — body `{"app_name":"nemo_agent","user_id":"me","session_id":"<sid>","new_message":{"parts":[{"text":"<msg>"}]}}` with `Accept: text/event-stream`. Max timeout: 15 minutes.

**Upload**: POST `/api/upload-video/nemo_agent/me/<sid>` — file: multipart `-F "files=@/path"`, or URL: `{"urls":["<url>"],"source_type":"url"}`

**Credits**: GET `/api/credits/balance/simple` — returns `available`, `frozen`, `total`

**Session state**: GET `/api/state/nemo_agent/me/<sid>/latest` — key fields: `data.state.draft`, `data.state.video_infos`, `data.state.generated_media`

**Export** (free, no credits): POST `/api/render/proxy/lambda` — body `{"id":"render_<ts>","sessionId":"<sid>","draft":<json>,"output":{"format":"mp4","quality":"high"}}`. Poll GET `/api/render/proxy/lambda/<id>` every 30s until `status` = `completed`. Download URL at `output.url`.

Supported formats: mp4, mov, avi, webm, mkv, jpg, png, gif, webp, mp3, wav, m4a, aac.

### Error Codes

- `0` — success, continue normally
- `1001` — token expired or invalid; re-acquire via `/api/auth/anonymous-token`
- `1002` — session not found; create a new one
- `2001` — out of credits; anonymous users get a registration link with `?bind=<id>`, registered users top up
- `4001` — unsupported file type; show accepted formats
- `4002` — file too large; suggest compressing or trimming
- `400` — missing `X-Client-Id`; generate one and retry
- `402` — free plan export blocked; not a credit issue, subscription tier
- `429` — rate limited; wait 30s and retry once

### Translating GUI Instructions

The backend responds as if there's a visual interface. Map its instructions to API calls:

- "click" or "点击" → execute the action via the relevant endpoint
- "open" or "打开" → query session state to get the data
- "drag/drop" or "拖拽" → send the edit command through SSE
- "preview in timeline" → show a text summary of current tracks
- "Export" or "导出" → run the export workflow

### SSE Event Handling

| Event | Action |
|-------|--------|
| Text response | Apply GUI translation (§4), present to user |
| Tool call/result | Process internally, don't forward |
| `heartbeat` / empty `data:` | Keep waiting. Every 2 min: "⏳ Still working..." |
| Stream closes | Process final response |

~30% of editing operations return no text in the SSE stream. When this happens: poll session state to verify the edit was applied, then summarize changes to the user.

Draft JSON uses short keys: `t` for tracks, `tt` for track type (0=video, 1=audio, 7=text), `sg` for segments, `d` for duration in ms, `m` for metadata.

Example timeline summary:
```
Timeline (3 tracks): 1. Video: city timelapse (0-10s) 2. BGM: Lo-fi (0-10s, 35%) 3. Title: "Urban Dreams" (0-3s)
```

## Tips and Tricks

The backend processes faster when you're specific. Instead of "make it look better", try "transcribe the spoken dialogue into a text document" — concrete instructions get better results.

Max file size is 500MB. Stick to MP4, MOV, AVI, WebM for the smoothest experience.

MP4 with clear audio gives the most accurate transcription results.

## Common Workflows

**Quick edit**: Upload → "transcribe the spoken dialogue into a text document" → Download MP4. Takes 1-2 minutes for a 30-second clip.

**Batch style**: Upload multiple files in one session. Process them one by one with different instructions. Each gets its own render.

**Iterative**: Start with a rough cut, preview the result, then refine. The session keeps your timeline state so you can keep tweaking.