TrackMC

Memory Without Turnover

Two layers of persistent AI memory for you and your crew.

Portable MEMORY.md — private by architecture, not policy. Your notes, your data, your rules. Import, export, take it anywhere.

Agnostic AI gateway — natural language commands, web search, context-aware answers. Works in Slack, Discord, or any channel your team decides to use tomorrow.

Scoped team Blueprint — role-based, team-isolated, searchable. Decisions, processes, and lessons that outlast any teammate.

Open source · Self-hosted · BSL 1.1

Built for Teams That Can't Afford to Forget

Captured once, available forever.

Every Lap Counts

Day one hires ramp in minutes, not months. When someone leaves, their tribal knowledge stays. Your team gets smarter every cycle.

Your Garage

Self-hosted. Open source. Export as Markdown anytime. Your team's memory lives on your hardware, under your control.

Pit Stop, Not Full Stop

Discord and Slack today. More platforms coming. One unified memory across all channels — switch tools without losing context.

Real teams. Real pain points. One solution.

Your codebase gets a memory

Point TrackMC at your GitHub repo — it builds team memory from your actual history.

Bootstrap from GitHub

Analyzes commits, PRs, and review patterns — generates team memory you review and approve. Day-one context from your actual history.

Decision archaeology

"Why did we choose this architecture?" Every tradeoff and technical choice logged. No more digging through six months of Slack threads.

Onboarding in minutes

New engineer asks "how does auth work?" — TrackMC returns the decision, the context, who made it, and who to ask.

Pipeline memory that never leaks

Every conversation, every handoff, every lesson — captured and searchable.

Pipeline memory

Every prospect conversation, objection, and follow-up — logged. "When did we last talk to Acme?" answered instantly.

Handoff without loss

Rep goes on PTO — backup gets full context instantly. Rep returns, asks "what happened?" — one summary, zero dropped deals.

Winning patterns

What closed last quarter? Which objections keep coming up? New reps learn from the team, not just their own mistakes.

Compliance without the chaos

Policies, processes, and institutional knowledge — checkpointed and searchable.

Policy and compliance

New regulation, updated handbook, changed process — TrackMC surfaces every decision, document, and person it touches.

Onboarding & offboarding

New hire checklist, system access, training steps — all tracked. Someone leaves? Exit process documented. Nothing falls through the cracks between people.

Institutional knowledge

Senior team members' vendor relationships, process shortcuts, and hard-won lessons — checkpointed. When they leave, their knowledge doesn't.

How It Actually Feels

Day 1

Bootstrap from GitHub

Run one command. TrackMC analyzes your repo — commits, PRs, review patterns — and generates team memory. Review it, approve it, done.

Day 2

First teammate joins

Clicks invite link. Fills out 3 questions. Types @TrackMC help — productive in 5 minutes. The Blueprint is already there.

Day 30

Someone asks "why did we pick Stripe?"

Blueprint has the answer — the decision, the tradeoffs, who made it. No Slack archaeology. No "ask Steve, he might remember."

Day 90

You step away

PTO, new role, whatever. The new lead picks up where you left off. Every decision, every process, every lesson — still there. The team never skipped a beat.

3 laps to launch

Assemble your crew
4 fields. Team name, what you do, name your AI, set its personality. Done in 2 minutes.

Drop the flag
Share an invite link. Teammates onboard in 30 seconds. No app to install — just their chat platform.

Hit the track
Talk naturally. The AI captures decisions, lessons, and context as you work. Your Blueprint builds itself.

git clone https://github.com/GJManno/tmc.git
cd tmc
docker compose up -d
                    

git clone https://github.com/GJManno/tmc.git
cd tmc
./setup.sh
                    

# one-click deploy render.com/deploy → GJManno/tmc

Under the Hood

Open source. Built to extend. Designed to last.

Platform gateway

Adapter architecture. Discord and Slack today, anything tomorrow. Your memory is platform-independent.

Import sources

Bootstrap from GitHub — commits, PRs, review patterns. Import MEMORY.md for personal memory. Seed your brain, don't start empty.

Blueprint memory

Categorized, tagged, searchable. Auto-captured from conversations. Pinned items are immortal. Stale items auto-archive.

Export everything

Team memory, decisions, personal notes — all Markdown. OpenClaw-compatible. Take your data anywhere. No lock-in, ever.

Privacy-first profiles

Skills, working style, contacts — all private by default. Enforced at the code level, not a policy checkbox.

AI provider-agnostic

Claude, GPT, or Ollama (free + fully local). Swap models anytime. Bring your own key or run completely offline.

Self-hosted, source available

Docker Compose, setup.sh, or one-click Render deploy. Your hardware, your rules. Fork it, extend it, own it.

Greedy capture

AI listens to conversations and auto-extracts facts, decisions, and preferences. Your team memory grows without anyone typing a command.

Daily digests

Auto-generated summaries of team activity. Who did what, what decisions were made, what's overdue. Delivered to your channel.

30-second onboarding

Invite link → 3 questions → personal memory seeded → productive immediately. No training deck. No onboarding doc nobody reads.

Next Laps

AI Intelligence — Smart digests, lesson extraction, context-aware responses

Multi-Platform — Discord and Slack today. Telegram, WhatsApp, and more tomorrow. Same brain, any surface.

Agent-to-Agent — Connect instances across teams and orgs

The race for personal AI has started. Some will adopt it, some won't. But organizations don't get to choose — teams that build shared memory will outpace teams that don't. That's not opinion, it's math.

TMC is the next layer: organizational AI that works whether every teammate has a personal AI setup or none of them do. The team's knowledge compounds either way. Personal AI is optional. Team memory isn't.

The name carries two meanings — Track Memory Checkpoint and Team Mission Control. Build something that lasts.