Atlaso

Atlaso is a memory layer that sits between you and every AI tool you use, so context stops being trapped inside whichever app you happened to be in. Explain a project's conventions to Claude Code and the same understanding is available the next time you open Cursor or Codex, without pasting a summary or maintaining a rules file in each tool.

How it connects. Atlaso attaches over the Model Context Protocol, installed as a skill or hook rather than a separate app. Setup is a single shell script followed by a browser authorisation; you then pick which tools to link. There are no API keys to paste and no config files to hand-edit. Tools that cannot run a terminal command connect through a hosted connector URL instead. Claude Code, Cursor, Codex, Claude Desktop, OpenCode and Antigravity are supported today, with ChatGPT and Meta AI listed as coming; in principle any MCP-capable client can attach.

Two scopes of memory. Personal memory travels with you across every connected tool — how you like work explained, which stack you default to, standing preferences. Project memory is pinned to a repository and deliberately does not bleed into other projects, so a decision made in one codebase does not resurface as advice in an unrelated one. What Atlaso captures is framed as decisions, watch-outs and open questions rather than raw transcripts.

Privacy handling. Traffic is encrypted in transit and at rest, access tokens are stored only as hashes, and secrets are redacted before anything is written. The vendor states memories are never sold and never used to train models, and that on the free plan memory is not sent to any LLM. Individual memories or the whole account can be deleted at any time.

Pricing. Free covers one device and one tool with basic capture. Pro is $10/month and lifts the device and tool limits while adding Ambient Memory, background enrichment, memory search and JSON export. Build is $25/month and adds a developer memory API. Published by Atlaso Labs Inc.