Turn AI coding
into reusable context.

FML captures agent sessions, repo history, reviews, and team setup, then packages that evidence into context for the next change.

~/acme — panopticon install

The model can
read your files.
It cannot know
your history.

Commercial codebases are not just files. They are decisions, reviews, reversions, incidents, naming conventions, business rules, and local workflows that explain why the files look the way they do.

Agents can read the current tree, but they do not know the path that got you there. So each task starts with missing context: the old PR, the rejected approach, the test that matters, the product rule, the teammate workflow that already works.

FML turns that scattered evidence into reusable context for the next engineer and the next agent run.

Getting started

~/your-repo
$ npm install -g @fml-inc/fml
✓ installed FML plugin
✓ registered Claude Code hooks
✓ Codex and Gemini scanners enabled
✓ Panopticon capture active
$ fml login
→ opening browser…
✓ signed in to acme-eng
Ready. Run your agents as usual.

Install once.

01

Hooks register in Claude Code, Codex, Gemini, Pi, and more. Sessions land in a local SQLite, with optional sync to your team.

Claude Code · /pr-review
> /pr-review
feat/fml-observe · 4 sessions · 15 commits
CRITICALSandbox REST routes not registered in http.ts
CRITICALShell injection in dbPath at register.ts:95
WARNINGSync token org identity dropped in panopticon_http.ts

Reviews with the why.

02

Run /pr-review on any branch. Grounded in the sessions that produced the change — so you see the why, not just the diff.

Claude Code · agent
> Recover the session I started on my laptop
Codex · 38m ago · 4 prompts
Last edit: app/login/page.tsx
Last prompt: “still overflows on iOS”

Ask FML from anywhere.

03

Claude Code, Slack, Telegram, terminal — one FML, same memory of your sessions, repo, and team.

Works with

ClaudeCodexGeminiCursorVisual Studio Code
ClaudeCodexGeminiCursorVisual Studio Code
Slack
Telegram

Your team's AI work,
every morning in Slack.

FML posts a daily digest to a channel — top sessions, cost outliers, stuck signals, and what's spreading on the team. Telegram works the same way.

#engineering· 12 members
fml
FMLAPP· 9:00 AM
Yesterday's AI work
12 sessions · 4h 18m active · $14
Top sessions
signup flow refactorClaude · 2h 14m
worker retry fixClaude · 1h 06m
login CSSCodex · 38m
Worth watching
1 session ran 6h with retries spiking — looks stuck
panopticon-review spreading: 4 adopters this week
View dashboard →

What you get

AI sessions and repo history, packaged for the next change.

Session replay

Every AI coding session becomes a timeline and summary: prompts, tool calls, file edits, model output, token cost, and what changed in the repo.

Workflow patterns

See how the team actually codes with AI: active users, repeated workflows, tool usage, cost patterns, stuck sessions, and setup gaps worth fixing.

Setup diffs

Compare skills, hooks, permission rules, MCP servers, local docs, and model choices across the org. Turn one person's advantage into a team standard.

Git and PR evidence

Ingest git history and GitHub PRs so FML knows what changed, why it changed, which decisions were rejected, and which files carry hidden context.

Task-ready context

Expose memory through the CLI, Slack, and MCP so agents receive the right summaries, facts, files, hooks, and rules before they start editing.

Review and outcome learning

PR reviews, reverted changes, incidents, and repeated fixes flow back into the knowledge base. The system learns from outcomes, not just prompts.

Pricing

Start locally. Add sync when you want shared context, setup diffs, and repo memory across the org.

Free
$0
Panopticon local capture
Local SQLite session store
Claude Code, Codex, Gemini CLI
Session timelines and AI summaries
Local search and cost tracking
TeamMost popular
$49/org/month

Everything free, plus:

FML cloud sync and dashboard
Org-wide session search
Slack bot and scheduled summaries
Team AI coding analysis
Context packages and setup diffs
Enterprise

For orgs that need the full platform at scale.

Everything in Team, plus:

Repo memory and task context rollout
SSO, roles, and access policy
Custom integrations
Data controls and audit trails
Rollout support for engineering teams

Questions

Panopticon captures AI coding sessions: prompts, tool calls, file operations, model responses, token counts, costs, and generated session summaries. Everything starts local in SQLite on the developer's machine. You control what syncs to FML.

FML ingests git history, GitHub PRs, reviews, session summaries, and team setup data so agents can ask what changed, why it changed, who touched it, what was rejected, and what context belongs in the next task.

FML supports Claude Code, Codex CLI, and Gemini CLI, with access through the CLI, Slack, and MCP. GitHub is the first source for repo memory; more workflow integrations can be connected as the rollout expands.

A developer can start locally in a few minutes: install Panopticon, register the coding-tool hooks, install FML, and link the org. Team rollouts add sync, dashboards, Slack, and shared context rules.

Better prompts help one session. FML captures the work, indexes repo history, compares setup patterns, and packages the right context at task time. The next prompt gets better because the codebase has memory.

Yes. Team analysis can surface skills, hooks, permission rules, MCP servers, local docs, repeated workflows, cost patterns, and stuck sessions. The product direction is to make the best patterns reviewable and adoptable instead of trapped on one laptop.

Reach out to us at hi@fml.inc and we'll get back to you.

Give the next change
the context it needs.