Product overview

One memory layer.
Many surfaces.

memry sits between your team's working surfaces and any human or agent that needs an answer. The graph, the retrieval, the trace, and the API — all in one place.

◆ Ingest

Index where work happens. Don't ask anyone to write more docs.

memry indexes the surfaces your team already uses, then keeps them in sync. No double-entry, no "remember to log it in memry."

Docs & wikis

Notion, Confluence, Google Drive, Coda, Outline. Live indexing — edits show up in retrieval within minutes.

Chat & meetings

Slack threads, MS Teams, Zoom transcripts, Granola, Fathom. Decisions get extracted from the chatter automatically.

Engineering surfaces

GitHub PRs, issues, design docs. Linear tickets and Jira epics. memry threads them into the graph for traceability.

CRM & ops

HubSpot, Salesforce, Pipedrive, Attio. Customer context lands in the same graph as engineering decisions.

Ingest API

A simple POST /ingest with optional source-stable IDs. Use it for internal tools, scripts, or data we don't natively support.

Permissioned by source

memry respects the ACLs of every source it indexes. If you can't see the Notion page, you can't retrieve it via memry.

◆ Retrieve

Ask in chat. Ask in your IDE. Ask through an agent.

Every retrieval returns a ranked list of facts with citations, the underlying graph path, and a confidence score. No black box.

Web app

Cmd-K search across the whole graph. Filters by author, source, project, and timeframe. Save searches as views.

Slack bot

@memry in any channel. Returns answers with the source thread linked. Replies inline so the answer stays where the question was asked.

CLI & IDE

memry CLI for terminal-native lookup. VS Code, Cursor, JetBrains plugins surface context next to the file you're editing.

◆ For agents

MCP-native. Drop-in for any agent stack.

memry exposes the graph as a Model Context Protocol server, so any MCP-aware client — Claude, Cursor, your in-house orchestrator — can use it as a tool. No glue code required.

Tool · search

Semantic retrieval

Natural-language query → ranked snippets with stable entity IDs and source URLs. Returns a confidence score so the agent can decide whether to ask a follow-up.

Tool · resolve

Entity lookup

Pass an entity ID, get the full subgraph. Useful when an agent is mid-task and needs to enumerate all known decisions about, say, "the Stripe webhook migration."

Tool · trace

Audit & explain

Returns the exact messages, docs, and edits behind an answer. Required for any compliance-sensitive use case — and for the human reading the agent's report.

◆ Security & compliance

EU-hosted. Built for regulated industries.

Encryption at rest & in transit

AES-256 at rest, TLS 1.3 in transit. Customer-managed keys available on the Scale plan.

Audit log

Every query, every ingest, every export — captured with actor, time, and scope. Available via SIEM webhook.

Self-host option

VPC-deployable for regulated buyers — banking, healthcare, public sector. Same code path, your network.

Bring memry into your team.

15-minute pilot setup. No commitment. Walk away with a real retrieval surface on your real data.