canonical: https://jentic.com/apis/nanoclip.ai/nanoclip

# NanoClip Public API

The NanoClip Public API uploads video or audio as a project and extracts transcript, speaker, face, scene, and retake intelligence from it. Processing runs as jobs: a request starts transcript, vision, or retake-removal analysis for a project, and a matching read returns the result once it is ready.

## For AI agents

Upload video or audio as a project and extract transcript, speaker, face, scene, and retake intelligence, starting each analysis as a job and fetching the result when ready.

## Scope

Does not handle video hosting, editing exports, or billing. Use for uploading media and extracting transcript, vision, and retake intelligence only.

## Capabilities

- Upload a video or audio file as a project for processing
- Create and retrieve a transcript with speaker information
- Run vision analysis to detect faces and scenes in a project
- Create and retrieve retake removal intelligence for a project
- Read project status and account details

## Use cases

### Video Transcription Inside an AI Agent

An AI agent uploads a recording as a project, starts a transcript job, and fetches the transcript with speaker information once it is ready. The create transcript and get transcript operations give the agent a two-step flow it can drive on its own, so a content assistant can turn raw footage into text without hosting a speech model.

Example prompt: Upload the media as a project, call create transcript, then poll get transcript until the result is ready.

### Face and Scene Analysis

A workflow analyzes the visual content of a clip to find faces and scene changes. The create vision operation starts the analysis for a project and the get vision operation returns the detections, giving an agent structured video intelligence to index or edit against.

Example prompt: Call create vision for the project, then read the faces and scenes from get vision.

### Retake Removal for Cleaner Cuts

An editing agent identifies retakes so it can assemble a cleaner cut. The create retake-removal operation starts the analysis and the get retake-removal operation returns the segments to drop, letting the agent propose an edit automatically.

Example prompt: Call create retake-removal for the project, then fetch the segments to remove from get retake-removal.

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| POST | `/v1/projects/upload` | Create an upload project |
| POST | `/v1/projects/{project_id}/transcript` | Start a transcript job |
| GET | `/v1/projects/{project_id}/transcript` | Get the transcript result |
| POST | `/v1/projects/{project_id}/vision` | Start face and scene analysis |
| POST | `/v1/projects/{project_id}/retake-removal` | Start retake removal analysis |

## Key resources

- **Projects** — Upload video or audio and read project status
- **Transcript** — Start a transcript job and fetch the speaker-labeled result
- **Vision** — Detect faces and scenes in an uploaded project
- **Retake removal** — Identify retake segments to drop from a recording

## Why Jentic

- **Setup:** Wiring the NanoClip Public API by hand means uploading a project, holding its bearer token, and coordinating the create-then-fetch calls for transcript, vision, and retake-removal results yourself. Through Jentic you install once, import it from the API Directory, store the token once, and your agent calls it.
- **Permission scoping:** Because operations are addressed by project ID, a rule can pin your agent to reading transcripts without letting it create new projects. You choose the operations and projects it may reach, so the rest stay out of scope unless you add them.
- **Credential handling:** Your NanoClip bearer token is stored once, encrypted, by your own Jentic One instance and injected into the Authorization header at execution time. It never enters the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'transcribe a video' or 'detect faces and scenes', and Jentic returns the matching NanoClip operation with its input schema so the agent calls the right endpoint without browsing the reference docs.

## Related APIs

- **AssemblyAI** — Speech-to-text with speaker labels and audio intelligence.
- **Deepgram** — Fast speech-to-text for audio and video.
- **Mux** — Video hosting, encoding, and delivery.

## FAQ

### What authentication does the NanoClip Public API use?

Every operation requires a bearer token in the Authorization header, per its OpenAPI spec. Through Jentic the token is stored encrypted in your own Jentic One instance and injected at call time.

### How do I get results back from the NanoClip Public API?

Processing follows a create-then-fetch pattern: a POST operation such as create transcript, create vision, or create retake-removal starts the job for a project, and the matching GET operation returns the result once it is ready. An agent creates the job and then polls the GET operation.

### What are the rate limits for the NanoClip Public API?

The OpenAPI spec does not specify rate limits. Check the NanoClip documentation at https://docs.nanoclip.ai for current limits, especially for large uploads and long processing jobs.

### How do I transcribe a video with the NanoClip Public API through Jentic?

Search Jentic for 'transcribe a video', add the NanoClip Public API from the directory, and your agent uploads a project, then starts a transcript job and fetches the result with your stored bearer token. To run it on your own infrastructure, install Jentic One from its GitHub repo.

### Is there a NanoClip Public API MCP server?

You don't need an MCP server to give your agent the NanoClip Public API. Jentic connects it directly from the API Directory: import it, store your bearer token once, and your agent calls the upload, transcript, and vision operations. Nothing extra loads into the agent's context until a call is made.

### Can I limit what my agent is allowed to do with the NanoClip Public API?

Yes. Write a rule that allows just the operations you want, such as fetching a transcript for one project without starting new vision jobs, so the agent cannot call anything else, and every call is logged. Because operations are scoped by project ID, rules can bound the exact projects your agent reaches.