# Google Video Intelligence API alternative

> Google deprecated the Cloud Video Intelligence API on 14 September 2026, and it shuts down on 14 September 2027. Video Context publishes this guide. Facts checked on Google’s pages on 1 October 2026.

## Feature map

| Google feature | What it returns | In Video Context (planned API design) |
|---|---|---|
| `LABEL_DETECTION` | Objects, scenes, and concepts | Labels and objects with timestamps, plus search in plain language |
| `SHOT_CHANGE_DETECTION` | Camera shot changes | Shots and scenes on the shared timeline |
| `SPEECH_TRANSCRIPTION` | Speech to text | Transcript with timestamps and speaker labels |
| `TEXT_DETECTION` | Text in the frames (OCR) | On-screen text with the times when it shows |
| `OBJECT_TRACKING` | Objects across frames | Objects and people with tracked positions |
| `LOGO_RECOGNITION` | Brand logos | Logos and brands with timestamps |
| `PERSON_DETECTION` | People and their attributes | People with timestamps and positions |
| `FACE_DETECTION` | Faces and attributes | Faces on request. Tell us your use case on the call. |
| `EXPLICIT_CONTENT_DETECTION` | Adult content | Content flags that you can route to review |

## Your options

- **Gemini (Google’s recommendation):** Flexible questions about each video. You write the prompts, and you build the index, the schema, and the search yourself. You pay per token for each request.
- **Azure AI Video Indexer or Amazon Rekognition:** Ready-made labels from another cloud. Rekognition closed streaming video analysis to new customers on 30 April 2026.
- **Video Context:** The same kinds of data in one schema, indexed once and searchable many times, with automations and agent tools. It runs in our cloud, your cloud account, or on your own servers. Private beta.

## Migrate in four steps

1. List the features that you call today, for example LABEL_DETECTION and SPEECH_TRANSCRIPTION.
2. Send us a few sample videos and one example of your current JSON output.
3. We map each field to the Video Context schema and send you the result for the same videos.
4. You compare the two results, then switch your code to the new endpoint.

Book a migration call: https://calendly.com/sam-reactvideoeditor/30min

Sources: [Google deprecation notice](https://docs.cloud.google.com/video-intelligence/docs/deprecations), [Google feature list](https://docs.cloud.google.com/video-intelligence/docs/features)

## FAQ

### When does the Google Video Intelligence API shut down?

Google deprecated the Video Intelligence API on 14 September 2026. It shuts down on 14 September 2027. After that date, calls to the API will stop.

### What does Google recommend instead of the Video Intelligence API?

Google recommends the Gemini family of models. Gemini answers prompts about video, but it does not give the same fixed feature output, so you must write prompts and build your own schema and index.

### Is there a drop-in replacement for the Video Intelligence API?

There is no exact drop-in replacement. Video Context gives the same kinds of data, such as labels, shots, speech, on-screen text, objects, and logos, in one stable JSON schema, and the team helps beta partners map their fields.

### Can I run the replacement in my own cloud?

Yes. Video Context can run in our cloud, in your cloud account, or on your own servers, so your video does not need to leave your network.

### Is the Video Context API available now?

Video Context is in private beta. Book a call, and the team helps you plan the migration.