> ## Documentation Index
> Fetch the complete documentation index at: https://docs.perceptron.inc/llms.txt
> Use this file to discover all available pages before exploring further.

# Services overview

> Learn how services differ from models and which service fits your task.

A service is an endpoint with its own request and response shape that does one specific job. You call it directly, rather than picking a model with the `model` field of a chat completions request.

Services and models use the same [Perceptron API key](https://docs.perceptron.inc/perceptron-mk1.5/guides/python/auth.md) and the same base URL, `https://api.perceptron.inc`. Send the key in the `Authorization: Bearer` header.

| Service | What it does | How it runs | Start here |
| - | - | - | - |
| Detect | Returns object detections for a single image, as boxes or points in image pixels. | One synchronous request. | [Detect](https://docs.perceptron.inc/services/detect.md) |
| Egocentric video annotation | Annotates first-person videos with hand tracking, hand activity, and a timeline. | Asynchronous batches that you poll, then download. | [Egocentric video annotation](https://docs.perceptron.inc/services/egocentric.md) and [limits and errors](https://docs.perceptron.inc/services/egocentric-limits-and-errors.md) |

The endpoint schemas are in the API reference: [Detect objects](https://docs.perceptron.inc/api-reference/detection/detect-objects.md) and [Submit an egocentric batch](https://docs.perceptron.inc/api-reference/egocentric/submit-an-egocentric-batch.md).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.