Introducing trellismem

Every team now works with AI agents, and every agent starts each task knowing nothing about the team. The contract terms are in Drive. The decision is in a Slack thread. The reason for the workaround is in a closed Jira issue. People know where to look. Agents do not, so people paste context into prompts all day.

trellismem is our answer: one memory that your team and its agents share.

What it is

trellismem reads the places your team’s knowledge already lives and builds a knowledge graph from them: the people, customers, projects and systems you talk about, and how they relate. You can ask it questions in the app. Your agents can ask it the same questions over MCP or a plain HTTP API, and they get the same answers.

A trellis is a lattice that gives a growing plant something to hold on to. That is the idea. Knowledge grows wherever work happens. trellismem gives it a structure.

What is in the first release

Sources. Connect Slack, GitHub, Linear, Notion, Google Drive, Gmail, Google Calendar, Confluence, Jira, HubSpot, Salesforce, Intercom and Stripe. Choose the channels, projects or folders to include, and memory stays in step with them.

Agents. Claude Code, Codex, Claude Desktop, Cursor, VS Code, Gemini CLI, Cline and more have guided setups in the app. Each agent gets its own API key.

Access control. Knowledge is organised into datasets, each with its own permissions and its own isolated storage. A question only ever searches what the asker may read.

Sign-in that fits a team. Passkeys, optional two-step sign-in with an authenticator app, and invite-only workspaces.

Nothing to install. trellismem runs in the cloud. Your team signs in from a browser, and agents connect with an API key.

Try it

Read the quickstart, or look at everything it does.