Memnex documentation¶
Pick the doc that matches what you're trying to do.
By role¶
| If you are... | Start with |
|---|---|
| Integrating an AI agent with Memnex (most common) | integration-guide.md |
| Running Memnex itself (self-hosting or as service operator) | operator-guide.md |
| Contributing to Memnex (extending, fixing, releasing) | ../CONTRIBUTING.md + publishing.md |
| Just exploring | The project README, then architecture.md |
Reference docs¶
| Doc | What's in it |
|---|---|
| architecture.md | System design, data flow, the four invariants. Read this to understand how Memnex works internally. |
| mcp.md | Full MCP API reference — 5 tools, 3 resources, 3 prompts, schemas, error shapes. |
| api-reference.md | REST API, Python embedded API, CLI commands. Audience-tagged so each section is clearly tenant- or operator-facing. |
| security.md | Multi-tenant isolation, audit ledger, regulated PII handling, injection floor. |
| features.md | What's built today and what's intentionally out of scope. |
| deployment.md | Docker compose, scaling, env vars. |
| eval.md | The six benchmark suites and how to run them. |
| comparison.md | How Memnex compares to Mem0, Zep, Letta, OpenAI memory. |
| roadmap.md | What's next and why, in three priority tiers. |
| publishing.md | How releases get to PyPI. Maintainer-facing. |
Suggested reading paths¶
First-time tenant integrating an agent: integration-guide.md → mcp.md → done.
Operator setting up self-hosted Memnex: operator-guide.md → deployment.md → security.md.
Engineer evaluating Memnex against alternatives: architecture.md → comparison.md → features.md → roadmap.md.
Maintainer cutting a release: ../CONTRIBUTING.md → publishing.md → ../CHANGELOG.md.