Decision layer for AI coding agents
Prove and defend the ROI of AI-assisted engineering.

Stop re-explaining your codebase to AI — every session, every tool.

Contexer captures your team's typed engineering decisions — with rationale — once, and replays them into every AI coding session. Reviewed by humans. Consistent across Claude Code, Cursor, Codex, and Gemini.

Stop paying, over and over, to re-establish context your team already has.

No account required · Decisions stored locally · Vendor-neutral across every major AI coding agent

Supported agents

  • /claude code
  • /cursor
  • /codex
  • /gemini cli
· One root cost, four symptoms

You keep paying to re-create knowledge your team already has.

Onboarding, re-explanation, alignment, and drift look like four separate problems. They're one: every AI coding session — and every new engineer — starts from zero on decisions your team already made.

01

Kill the AI Re-Explanation Tax

Every session starts with your rules, constraints, and architecture already loaded — so you stop paying the model, in tokens and time, to re-learn what your team already decided.

02

Onboard humans AND agents instantly

A human onboards once. An AI agent re-onboards cold on every session. Contexer is onboarding-for-agents, every time — decisions and rationale on tap, whether the reader is a new hire or a fresh Claude window.

03

Align the team without the meeting

One reviewed source of engineering decisions, shared across everyone's AI tools. Knowledge spreads without alignment syncs, and nobody re-agrees on things you already agreed on.

04

Stop architecture drift

AI can't quietly violate decisions it can see. Approved decisions are enforced across every tool, so drift shows up as a reviewable proposal — not a merged PR three sprints later.

· Workflow preview

Capture, propose, approve. Right inside your terminal.

These mockups show the intended flow. The actual UI may differ. The idea is decisions captured, reviewed, and replayed where your AI coding assistant already lives.

01Capture
Live
>
Captured as global convention
Rule
Release sequence: branch → bump → PR → merge → tag
Loaded into 0 sessions · applies to all future repos

State a rule. Contexer stores it.

Directives you state outright become global conventions, loaded into every future session.

02Propose
Idle
Pending decisionAI proposed
Use RabbitMQ for async jobs, not Kafka.
Rationale: lower ops overhead, fits current scale.
1 approve
2 edit
3 ignore

AI surfaces a decision mid-work.

Architecture and constraint decisions are parked for review, never replayed until you approve.

03Approve
Idle
>
Approved · v1
Decision
Use RabbitMQ for async jobs.
immutable
replayed across agents

One keystroke. Versioned forever.

Approve, edit, or ignore. Approved decisions are immutable and replayed across every agent.

· What Contexer actually captures

Engineering decisions. Not chat history.

Architecture Decisions

"We standardized on PostgreSQL."

Constraints

"Never expose S3 buckets publicly."

Conventions

"Terraform modules live under infrastructure/modules."

Patterns

"Every API emits OpenTelemetry traces."

These are engineering decisions - not chat history, not transcripts, not session logs.

· Not another AI memory tool

AI memory tools recall conversations. Contexer is your team's decision of record.

CLAUDE.md, AGENTS.md, .cursorrules, and per-assistant memory features are single-player: one person edits a file, another tool ignores it, nothing is reviewed, nothing is versioned. Contexer is the shared layer above them — typed engineering decisions, human-approved, versioned forever, and replayed into every agent your team uses.

AI memory / rules files
Contexer
What it holds
Chat history, per-assistant rules files
Typed engineering decisions with rationale — the team's decision of record
How it gets in
Auto-scraped from sessions, or hand-edited by one person
AI proposes, a human reviews and approves before it's trusted
Versioning
Overwritten in place, no history
Approved decisions are immutable; changes create a new revision with full provenance
Scope
Per-assistant, per-machine
One decision layer, shared across the team
Vendor coverage
Locked to one tool at a time
Vendor-neutral: Claude Code, Cursor, Codex, and Gemini CLI
What you can defend
"The model saw some notes."
Who approved what, when, and why — auditable end to end
· Prove and defend your AI ROI

Governance is how you keep your AI budget.

Every decision in Contexer is typed, human-approved, versioned forever, and traceable to the session and person that produced it. When finance, security, or the board asks what you're actually getting for your AI spend, you have the receipt: an auditable trail of the decisions your agents are following, who approved them, and why.

It's what teams reach for the moment AI budgets get scrutinized.

01

Reviewed, not scraped

AI-proposed decisions wait for a human before they're trusted. Nothing enters the team's decision layer without an approver.

02

Immutable and versioned

Every approved decision has provenance and history. You can answer who decided what, when, and why — down to the session it came from.

03

Vendor-neutral by design

One decision layer across Claude Code, Cursor, Codex, and Gemini. Your governance isn't locked to whichever model wins next quarter.

· Pricing

Start free. Upgrade when your team grows.

Most popular

Free

Freeforever

The full decision layer, available two ways. Run the local CLI and JSON store with no account, or sign in for managed cloud sync and the web dashboard.

MIT licensed · No account required for local OSS

Open Source CLI

Local JSON store. Fully offline. No web app or cloud. No account required.

Personal Cloud

Encrypted sync, web dashboard, and cross-machine replay.

  • Capture architecture, constraint, convention, and pattern decisions
  • AI-proposed decisions held for your review
  • Approve, edit, skip, or dismiss before replay
  • Injected into every Claude Code, Cursor, Codex, and Gemini CLI session
  • Web dashboard to review history and manage decisions (Cloud)
  • One-click upgrade to Teams

Teams

Freeduring early access

One repository per team. Invite members and share your approved engineering decisions across everyone's AI coding sessions — one decision of record, every agent. Pricing will be announced once our IP process is settled.

No credit card required · Pricing coming soon

  • Everything in Free Cloud
  • Unlimited team members with OAuth sign-in
  • One shared repository per team
  • Shared decision repository via team MCP endpoint
  • Role-based approval workflow (lead / member)
  • Confidence scores with supporting evidence - leads only
  • AI-generated canonical suggestions and updates
Contact us

Organization

Custom

For orgs running multiple teams. One decision layer across every team, with enforcement, drift detection in CI, SSO, and an auditable trail of the AI spend you're defending.

SSO · Audit trail · Pipeline drift detection

  • Everything in Teams
  • Cross-team decision sharing across the org
  • Enforcement policies (require approval, block on drift)
  • CI/CD pipeline hook - detects and rates architecture drift on every PR
  • SSO login (SAML, OIDC)
  • Full audit trail and access controls
  • Dedicated support and onboarding
· Quick start

Two commands. Under two minutes.

Requires Python 3.12+ and uv. Auto-detects Claude Code (~/.claude), Cursor (~/.cursor), Codex (~/.codex), and Gemini CLI (~/.gemini) and wires them in.

bash · install
# Step 1. install
$ uv tool install contexer

# Step 2. wire into your AI assistant
$ contexer install

-> detected ~/.claude (Claude Code)
-> detected ~/.cursor (Cursor)
-> detected ~/.codex (Codex)
-> detected ~/.gemini (Gemini CLI)
-> wired all four. restart your AI assistant.

Restart your AI assistant and open any git repo. Contexer runs from there.

· What the decision layer costs at runtime

~26 tokens per rule. Paid once per session.

Approved constraints and conventions are the only thing replayed at session start. Architecture and pattern decisions cost nothing until the work actually needs them, when the agent pulls them on demand. Lookups are sub-millisecond and run before the response is generated — a fraction of the tokens you were spending to re-explain the same decisions by hand.

Pre-loaded rules
Tokens at session start
5 rules
~125
10 rules
~250
25 rules
~625

Deduplication is deterministic - >70% token overlap is dropped. No LLM call. No tokens.

session start context
## Project rules - apply to ALL tasks in this repo:
- [convention] Use uv, not pip, for
  all dependency management
- [constraint] Never commit untested
  code - CI blocks merges below coverage

2 architecture decision(s) stored. Call
get_context before reading files for
questions about architecture or rationale.