Skip to main content

​
What is 3ngram?

3ngram gives all your AI tools one shared brain. It provides persistent memory, searchable knowledge, and proactive follow-through across every AI tool you use. Think of it as a brain for what your AI knows and a chief of staff for what it does with that knowledge.

​
Who is it for?

3ngram is for anyone who does real work inside AI tools and needs that work to carry forward. Consultants managing multiple clients, product managers coordinating across teams, founders tracking decisions across every function, and developers working across multiple AI tools. AI power users are a strong secondary audience, especially for advanced MCP workflows.

​
How does it connect to my AI tools?

3ngram uses the Model Context Protocol (MCP), an open standard for connecting AI assistants to external tools and data. See Connecting Clients for setup instructions.

​
Which AI clients are supported?

Any MCP-compatible client works. Tested with:
  • Claude Desktop
  • Claude Code CLI
  • Cursor
  • ChatGPT

​
Is there a free tier?

Yes. The Starter plan is free with no credit card required. Paid plans unlock unlimited memories and content sources. See Plans and Billing for the full comparison.

​
Where is my data stored?

PostgreSQL on managed cloud infrastructure. Data is encrypted in transit and at rest.

​
Can I export my data?

Yes, at any time, regardless of subscription status. Go to Settings > Export to download a ZIP of all your documents and memories.

​
Is my data sent to LLMs?

Document and memory text is sent to OpenAI for embedding generation (semantic search). Memory consolidation via LLM synthesis is opt-in and defaults to off. See our security page for details.

​
How does 3ngram handle privacy?

  • Row-level security enforces strict user isolation at the database layer
  • Audit trail of admin actions
  • <private> tags let you exclude sensitive content from storage

​
Are team features available now?

Team features are planned. Check the release notes for current availability.

​
ChatGPT and Claude already have memory. Why do I need 3ngram?

Native AI memory stores preferences and facts. It does not track commitments, does not know deadlines, does not proactively remind you of anything, and is locked to one provider. 3ngram adds structured memory types (decisions, commitments, blockers), cross-provider continuity, and an accountability layer that turns memory into follow-through.

​
Can I just use a notes app instead?

Notes do not transfer into your AI’s context. You would have to copy-paste your notes into every AI session, manually keep them updated, and remember to reference them. 3ngram makes your knowledge automatically available inside every AI tool you use, with semantic search and lifecycle tracking built in.

​
Is MCP too new or niche?

MCP is supported by Claude, ChatGPT, Cursor, Windsurf, and dozens more tools. It is governed by the Linux Foundation and backed by both Anthropic and OpenAI. Adoption is growing fast. MCP is becoming the standard protocol for connecting AI tools to external data and capabilities.

​
Why not build this myself?

You can. 3ngram offers a free Starter tier so you can try it without commitment. But lifecycle semantics (multiple memory types with status, deadlines, recurrence, resolution), integrations, MCP tools, prompt templates, and resources, RAG search, knowledge graph, scheduled agents, event triggers, and a web dashboard represent months of engineering. Most people want the finished product.

​
How is this different from CLAUDE.md?

CLAUDE.md is a static file in one repo. It does not update itself, does not work across providers, cannot track deadlines, and does not proactively surface anything. 3ngram is a living system that builds knowledge over time, works across every AI tool, and acts on what it knows.

​
How does 3ngram compare to other tools?

See our detailed comparison pages: