Skip to content

feat: Distribution & Adoption Engine — framework adapters, comparisons, install script, README#205

Merged
tcconnally merged 7 commits into
mainfrom
feat/distribution-adoption-engine
Jun 22, 2026
Merged

feat: Distribution & Adoption Engine — framework adapters, comparisons, install script, README#205
tcconnally merged 7 commits into
mainfrom
feat/distribution-adoption-engine

Conversation

@tcconnally

Copy link
Copy Markdown
Collaborator

Distribution & Adoption Engine for Mimir

Framework Adapters (critical fix)

  • All three adapters: Added "serve" subcommand to args — mimir --db <path> without serve exits with help text. The fake Popen in tests masked this.
  • LangGraph: Fixed init-handshake — now captures init_id and consumes response in _ensure_session (matching crewai/autogen pattern)

Comparison Pages

  • docs/comparison/mimir-vs-mem0.md — local-first vs cloud-only, honest strengths and weaknesses
  • docs/comparison/mimir-vs-letta.md — memory engine vs agent runtime, composability
  • docs/comparison/mimir-vs-zep.md — binary vs infrastructure, MCP-native advantage

README Overhaul

  • Hero tagline: "Persistent Memory for AI Agents — MCP-Native. Local-First. Zero Dependencies."
  • One-line install + 30-second quickstart
  • Framework integration badges
  • Full comparison matrix with Mem0, Letta, Zep
  • Stress test results: 100K entities in 1.01s, recall 0.022s, decay 1.317s
  • All 36 MCP tools documented

Distribution

  • scripts/install.sh — one-line install: curl -sSf ... | sh (Linux, macOS, WSL)
  • Dockerfile CMD fixed to include serve

Discovery

  • awesome-mimir.md — curated integrations, tools, comparison matrix

Glama + Smithery

  • Updated server.json (v2.0.0, 36 tools), glama.json, smithery.yaml (fixed args)

Test Results

  • Rust: 32/32 passing (stress test: 100K entities in 2.41s)
  • Python adapters: 16/16 passing (autogen 7, crewai 3, langgraph 6)

…handshake

All three adapters (LangGraph, CrewAI, AutoGen) were using
`mimir --db <path>` without the required `serve` subcommand,
which caused Mimir to print help and exit immediately.

LangGraph adapter also fixed: init_id is now captured via _next_id()
pattern and the init response is consumed in _ensure_session,
matching the crewai/autogen patterns.

Without these fixes, the real Mimir binary would fail to start
(the fake Popen in tests masked the issue).
Honest, detailed comparisons highlighting Mimir's unique positioning:
- MCP-native, local-first, zero-dependency
- Each page covers architecture, memory models, MCP tools, and
  'when to use which' guidance
- Strengths AND weaknesses candidly assessed
…rs, 36 tools

- New hero: 'Persistent Memory for AI Agents — MCP-Native. Local-First. Zero Dependencies.'
- One-line install + 30-second quickstart
- Framework integration badges (LangGraph, CrewAI, AutoGen)
- Full comparison matrix with Mem0, Letta, Zep
- Stress test results: 100K entities in 1.01s, recall 0.022s, decay 1.317s
- All 36 MCP tools documented with descriptions
- Updated CLI flags, transport options, and Perseus config
Supports Linux (x86_64, aarch64), macOS (x86_64, aarch64), WSL.
Auto-detects platform, downloads binary from GitHub releases,
installs to ~/.local/bin, verifies, and prints quickstart.
`mimir --db /path` without `serve` exits immediately with help text.
Fixed to `mimir serve --db /data/mimir.db`.
Community discovery page with framework integrations, MCP hosts,
comparison matrix, and contribution guide for listing projects.
Supports Linux (x86_64, aarch64), macOS (x86_64, aarch64), WSL.
Auto-detects platform, downloads binary from GitHub releases,
installs to ~/.local/bin, verifies, and prints quickstart.
@tcconnally tcconnally force-pushed the feat/distribution-adoption-engine branch from 18ad5d3 to b74d47f Compare June 22, 2026 01:46
@tcconnally tcconnally merged commit b74d47f into main Jun 22, 2026
6 of 7 checks passed
@tcconnally tcconnally deleted the feat/distribution-adoption-engine branch July 2, 2026 17:13
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant