# Threadless AI

> Hosted memory for AI agents. Before every model call your agent hands over what is new in the conversation and gets back the request to submit to OpenAI (Chat Completions or Responses); after the reply it reports it. Old episodes are forgotten by time, their facts are kept, and recall brings them back when the conversation turns to them.

If you are an agent adding Threadless AI to a codebase, read https://threadlessai.com/setup.md first and follow it step by step. Every page below is markdown.

## Setup

- [Setup for agents](https://threadlessai.com/setup.md): Exact steps for an agent adding Threadless AI to a codebase that calls OpenAI, from install to one verified call.

## Start

- [Overview](https://threadlessai.com/docs.md): What Threadless AI does, the three calls it takes, and where to start.
- [Quickstart](https://threadlessai.com/docs/quickstart.md): Install the SDK, create a memory, and run an agent loop through decorate and report, in TypeScript or Python, for Chat Completions or Responses.
- [The three calls](https://threadlessai.com/docs/calls.md): decorate before every model call, report after it, and the note tool the model writes facts with, answered with NOTE_RESULT.
- [Memory ids](https://threadlessai.com/docs/memory-ids.md): One memory per conversation. A memory only grows, so a branch or an edit is a new memory.
- [Models and formats](https://threadlessai.com/docs/models-and-formats.md): Chat Completions and Responses, which models work, where our messages go in the request, and what the size ceiling assumes.

## Behaviour

- [Failure behaviour](https://threadlessai.com/docs/failure-behaviour.md): When we cannot be reached, the SDK builds the request from a local tail of recent turns and your model call goes ahead. Memory never fails the model call.
- [Limits](https://threadlessai.com/docs/limits.md): Request size, request rate, metadata, memory ids, the size ceiling and the free plan's allowances.
- [Concepts](https://threadlessai.com/docs/concepts.md): Episodes, encoding, traces, activation, fading, sleep, significance and recall, named for the human memory they mimic.

## Reference

- [API reference](https://threadlessai.com/docs/api-reference.md): Every operation and schema of the runtime API, generated from openapi/runtime.yaml.
- [Admin MCP setup](https://threadlessai.com/docs/admin-mcp.md): Connect Claude Code, Cursor or Codex to the admin MCP server so your coding agent can create projects and API keys.

## Optional

- [OpenAPI spec, runtime API](https://threadlessai.com/openapi/runtime.yaml): the HTTP API as OpenAPI 3.1 YAML.
- [All docs in one file](https://threadlessai.com/llms-full.txt): the setup page and every docs page, concatenated.