# Praecise

> An AI lab, and the cloud to run it on. Praecise publishes open-source systems for building and running AI — a framework for agents and an acceleration layer for inference — and operates the cloud that deploys them: apps, agents, models, data and machines.

Praecise is the family name. Praecise Harness makes an agent work. Praecise Engine makes a model faster. Praecise Cloud is the managed side of an app written with the Harness.

## Open source (Apache-2.0)

- [Praecise Harness](https://github.com/Praecise/praecise-harness): The framework for AI agents. A folder is an app. TypeScript, Node 22+, npm package `praecise`. Serves chat UI, REST, MCP, A2A and AG-UI from one file; any MCP server or OpenAPI document becomes tools; durable workflows with approval gates; SQL, document, time-series and vector stores.
- [Praecise Engine](https://github.com/Praecise/praecise-engine): A backend-agnostic acceleration layer for LLM inference. Rust with a C ABI. Speculation policy, n-gram self-speculation without a draft model, licence-checked drafter catalogue, request shaping, admission control. llama.cpp is the first backend; verified on NVIDIA GB10.

## Start

```
npx praecise init my-app
cd my-app && npm install
npx praecise dev
```

Put `PRAECISE_API_KEY` in `.env` to run on Praecise Cloud, or describe your own model endpoints in `praecise.config.ts`.

## Cloud

Runs apps, agents, models, data, machines and automations on hardware Praecise owns and operates. Projects are deployed, configured and watched from the Console at https://console.praecise.com (sign-in by email code). Early workloads are placed by conversation: contact@praecise.com.

## Pages

- [Lab](https://praecise.com/lab): the two open-source systems, their metadata and features.
- [Cloud](https://praecise.com/cloud): what the platform runs and how a deploy works.
- [Research](https://praecise.com/research): measured notes from the Engine, negative results included.
- [About](https://praecise.com/about)
- [Contact](https://praecise.com/contact)