Mnemocompany memory

Operating memory for teams and AI agents

The memory of how your company actually works

Mnemo learns from the work your team already does in Slack, email, Jira, and GitHub. It keeps the decisions, exceptions, and lessons behind that work, and how they change over time, so new hires, handovers, and AI agents get answers with the source attached.

Under the hood: hybrid retrieval across seven signals (vector, BM25, atomic fact, temporal, entity graph, concept boost, semantic bridge) fused with intent-aware Reciprocal Rank Fusion, atomic fact extraction, and prompt-ready profiles. Model-agnostic and US-hosted. Free tier: 1K writes and 10K searches each month, no credit card.

Measured, not asserted

  • 20 — tools Mnemo learns from, backfilled and kept in sync.
  • 99.2% — evidence found in the top 5 results on LongMemEval-S (496 of 500, Recall@5, top-50 candidate chunks, no LLM reranker).
  • ~150ms — warm search p50 on the production API.

Most of how your company works was never written down.

It lives in Slack threads, ticket comments, and the heads of the people who were in the room. The wiki goes stale, search surfaces the old answer, and when someone moves on, the reasons go with them.

Wikis and docs

Written once, rarely updated. By the time someone needs the page, the process it describes has already moved on.

Search across tools

Finds every message that mentions the topic, including the decision that was reversed last month. It can't tell you which one still holds.

Mnemo

Keeps the decision, the exception, and whatever replaced it, tied to the thread or ticket where it happened, and ranks what's current higher.

From everyday work to answers you can check

I. Connect your tools

Link Slack, email, Jira, GitHub, and more. Mnemo backfills the history and keeps syncing.

II. Keep what matters

Mnemo pulls the decisions, exceptions, people, and dates out of threads and tickets, and keeps a reference to where each one came from.

III. Track what changed

When a newer decision replaces an older one, Mnemo marks the old one as superseded and ranks it lower. It stays on record, so you can see what it was and when it changed.

IV. Ask, and check the source

People ask in Slack, agents ask through the API or MCP, and answers come back with the message, ticket, or doc behind them.

Your team's memory, already filled in

Nobody has to write anything down. Connect the tools your team already works in and Mnemo keeps that context continuously in sync — 20 sources, backfilled and updated, all answerable from one place.

Built for the moments knowledge usually gets lost

The answer already exists somewhere in your tools. Mnemo helps the next person, or the next agent, find it.

  • Handovers — when someone changes roles or leaves, the next person can ask why things are set up the way they are and get the thread where it was decided.
  • New hires — learn how things are actually done here, including the exceptions nobody wrote down, without waiting for the one person who knows.
  • AI agents — give agents the same memory, scoped to what each one should see, so they can tell this quarter's rules from last quarter's.

Answers you can check, not just trust

  • Answers show their source — each result points back to the message, ticket, or doc it came from, so anyone can check it before acting on it.
  • What's current ranks higher — a decision that was replaced is marked superseded and sinks in results. It's kept, not erased, so you can still trace how things got here.
  • Ask what changed, and when — ask how a process worked before the reorg, or what changed since last quarter. Dates and order are kept, not flattened.
  • Small facts don't disappear — long threads become short, reusable facts without losing the people, dates, or source they came from.
  • Find the detail, not just a similar paragraph — Mnemo searches meaning, wording, facts, entities, and time together, then combines the evidence into one ranked result.
  • Limit what each agent can see — connect an agent over MCP and limit it to the containers it needs. Workspace isolation, scoped API keys, audit history, and US data residency are built in.

Find the right memory. Not just a plausible one.

98.0% recall with only five chunks. Increasing the candidate pool tenfold moves recall by 1.2 points — the result saturates early instead of depending on an oversized search pool.

Give your agents the same memory.

What your team's memory knows is one API call away. Write to it, search it, and get a reference to the source back with results.

When memory goes wrong, you should be able to see it.

Once people and agents rely on it, company memory is infrastructure. Mnemo gives your team the health, usage, audit, and delivery history needed to work out what happened without reconstructing it from logs across three systems.

Use the tools already in your stack.

Seven packages published across npm and PyPI, so the memory your connectors fill is one import away from any agent you build.

The packages: getmnemo, getmnemo-cli, getmnemo-mcp, getmnemo-anthropic, getmnemo-vercel-ai, getmnemo-mastra, and the Python getmnemo on PyPI. Install the SDK with `npm install getmnemo`, the CLI with `npm install -g getmnemo-cli`, or run the MCP server with `npx -y getmnemo-mcp`.

Keep what your team knows

Connect the tools your team works in, ask your first question, and see the source behind the answer. Free to start, no card.

Explore Mnemo