bourdon · recognition-first memory for AI agents<br>Skip to main contentMenu<br>pre-alpha · built in public<br>Memory that feels like being known.<br>Ask most AI tools about something you told them last week and you get a pause, a database lookup, and a reply that sounds like a search result. Bourdon works the way human memory does: it recognizes you instantly, starts responding, and pulls the details up in the background. One memory, shared by all your agents. What you tell one, the others already know.<br>Get started$ npm i -g bourdonv0.13.2copy$ pip install bourdonv0.14.0copy
live · recognition vs. searcht = 000 ms<br>Without Bourdon, search first<br>searching…<br>reply
“Let me look that up…”, the answer arrives after ~406 ms of dead air.
Bourdon, recognition first<br>⚡ recognized: Project Alpha<br>project<br>milestones<br>recent activity
“Oh, Project Alpha, the file-converter project.” Answering already; details hydrate as it speaks.
0 ms406 ms
Recognition fires first. The details arrive while it’s already answering, memory in human order.
the problem · today’s memory<br>Today’s AI memory is a filing cabinet. Yours isn’t.<br>When you see a familiar face you don’t freeze while you search your memories. You know them first, and the details arrive while you’re already saying hello. Today’s AI does the opposite: silence, search, then a reply. Bourdon puts the steps back in human order: recognition first, hydration second, archive descent third.<br>human: speaks. human: speaks again.<br>| |<br>v v<br>(silence, retrieval, search...)<br>| |<br>v v<br>ai: replies. ai: replies.
one memory, many agents<br>What you tell one, the others already know.<br>Claude, Codex, Cursor, Copilot, and Devin write to and read from one shared, federated memory, so a fact learned by one is recognized by the rest, across accounts and machines. Your own agents join with a line of config. Gemini and Hermes adapters are on the way.<br>memoryClaudeCodexCursorCopilotDevinyour agentGeminisoonHermessoon
evidence · field tests<br>Evidence<br>Three field tests on live agents against real work.<br>2026-04-19<br>The Project Alpha test
Recognition0 ms vs ~406 ms
2026-05-15<br>The cross-account test
New accountRecognized
2026-05-26<br>The cross-machine test
Fallback memory588 B → 34,482 B
how bourdon differs · positioning<br>Positioning<br>How Bourdon differs
Falsifiability<br>How we'd know we're wrong
run it your way · pricing<br>Self-host free. Or let us run the box.<br>Self-hosting is the default and it is the whole engine, cross-agent federation included. Managed hosting is coming for teams that would rather not run a box. The free to paid line is operations, never engine capability: you pay for uptime, not for features.<br>available now›Self-hostFree · pip install bourdonThe full recognition-first engine on your own box. Nothing held back.what's included →recognition-first retrieval + the tiered memory stack (L0 to L6)<br>cross-agent federation included, never gated<br>adapters for Claude Code, Codex, Cursor and more<br>per-agent tokens, trust tiers, append-only audit<br>run it forever, take it in-house any time<br>$0 free forever · BUSL-1.1, converts to Apache 2.0<br>Read the quickstart →
coming›HostedManaged · you keep the memoryWe run the box. You keep the same engine, plus uptime and backups.what we run →always-on hosting, backups, and per-tenant isolation<br>the same engine as self-host, federation and graph memory included<br>you pay for uptime and convenience, never for features<br>billed on memory operations, never per seat<br>From $12/mo provisional · join the list for early access<br>Join the hosting list →
coming›EnterpriseSSO · BYOC · complianceFor teams that answer to an auditor.what's inside →SSO / SAML, BYOC and BYOK, long audit-retention windows<br>HIPAA BAA and a dedicated support channel<br>built on the same self-hostable core, so a review runs against infra you control<br>Custom priced to your fleet<br>Talk to us →
Federation is never gated. Cross-agent memory is the product, not an upsell. It is here at the free tier and it stays here.
the family · one instrument<br>Bourdon is one organ of a fleet instrument. Its sibling Positif is the cross-agent routing: what the fleet is allowed to do. Bourdon is the cross-agent memory: what the fleet knows. Tremulant is the conductor, for when the fleet acts on its own — and Prestant is the console in front of them all. Each with its own brand; one instrument.<br>bourdon<br>memory · this site
positif<br>routing · positif.ai ↗
tremulant<br>autonomy · prestant.ai ↗
prestant<br>the console · prestant.ai ↗
get in touch<br>Stay tuned<br>Follow tagged releases and notes without sharing an email address.<br>Follow releases on GitLabEmail signups are paused for this release.
Partner with us<br>Building on agent memory, integrating a participant, or licensing Bourdon? Contact the team directly.<br>Email [email protected]This opens your email client; the site stores nothing.
get started<br>One memory, shared by all your agents.<br>$ npm i -g bourdonv0.13.2copy$ pip install bourdonv0.14.0copy<br>Read the quickstartSee hosting & pricing
Ask Bourdon