cedr.dev is in public beta. Expect fast changes and rough edges. Read the docs

Every coding agent, one control panel.

Route the work to the right agent, review it across vendors, keep every agent's setup in step, and see what it all costs. Free CLI, shared in the app.

Tell your agent:Read https://cedr.dev/docs/agents and help me set up cedr. If I already set products up in the cedr web app, use those settings. Otherwise act as my setup wizard: ask which products I want and whether to use Local MCP, Hosted MCP, both or neither, then walk me through each one a step at a time. Wait for my approval and configure only what I approve. Never ask for credentials in chat.
The cedr app's Monitor page: a week's spend at API prices, tokens, cache hit rate and plans in use, tokens by agent, each plan's limit, and the latest route picks

Built for every major coding agent

Turn your agents into a team

The cedr suite gives every agent you pay for the same setup, the same rules and one view of what is left. Here is what each tool does.

Route

Beta

Send each job to the right agent.

Route picks the agent, model and effort for each kind of work, such as a code review or deep research, then sends it to the plan with the most room left.

Read the Route Guide
$ cedr route pick review
target           codex/sol  effort high
reason           plan 64% left; claude at 91%, skipped
tag              rt_7Kq2

Works withClaude CodeCodexAntigravityMuse CodeT3 Code

Council

Beta

Several vendors review every risky change.

Council has reviewers from several vendors check the same change on their own, then merges their reports into one gate: pass, block or incomplete.

Read the Council Guide
$ cedr council plan review.council
codex/sol        high   room 64%
claude/opus      high   room 22%
antigravity/pro  high   room 80%
quorum           2: 3 eligible, run r-5c1e0a2b

Works withCodexClaude CodeAntigravity

Guard

Beta

Watch the skills and MCP servers you depend on.

Guard pins every skill and MCP server your agents load from a repo, records what each server offers a model, and tells you when any of it changes.

Read the Guard Guide
$ cedr guard outdated
github-mcp       1.4.0 -> 1.6.2
+ tool           create_workflow_dispatch
+ host           uploads.github.com
risk             high: new host

Works withClaude CodeCodexCursorGemini CLICopilot CLIopencodeAntigravity

Sync

Beta

One agent setup for every harness.

Sync keeps one setup for every coding agent: write rules, skills, MCP servers, permissions and hooks once in .cedr/, and cedr sync writes each agent's own files.

Read the Sync Guide
$ cedr sync
wrote            .mcp.json  (Claude Code)
merged           .codex/config.toml  (Codex)
wrote            .cursor/rules/20-testing.mdc
merged           .gemini/settings.json

Works withClaude CodeCodexCursorGemini CLICopilot CLIopencodeAntigravity

Monitor

Beta

Every agent's usage, spend and limits in one place.

Monitor reads each agent's local logs and plan readings: tokens, spend at API prices, cache hits, sessions, errors and how close each plan is to its limit.

Read the Monitor Guide
$ cedr monitor limits
Claude Code      5-hour   64%  tight
Codex            weekly   38%  on track
Muse Code        weekly   12%  on track

Works withClaude CodeCodexGemini CLICopilot CLIopencodeMuse CodeAntigravity

Bench

Beta

Know a change helps before it merges.

Bench runs real tasks against a baseline and a candidate, in your CI or on your machine and within a budget you set. Your tests decide each trial, and it reports the pass rates, the cost and how far the difference can be trusted.

Read the Bench Guide
$ cedr bench run
baseline         2/3 passed  $0.12
candidate        3/3 passed  $0.09
pass rate        +33 points, intervals overlap
spent            $0.21 of $2.00

Works withClaude CodeCodex

Everything a team running several coding agents needs, in one CLI and one app.

  • Plan balancing
  • Cross-vendor review
  • One setup, every agent
  • Drift checks in CI
  • Spend at API prices
  • Plan limit alerts
  • MCP server pinning
  • Org rules
  • Team budgets
  • SAML SSO
  • Audit export
  • Works offline
  • GitHub Action
  • Activity ladders

Pay only for the hosted app

Every CLI is free forever. app.cedr.dev is optional, and you pay per developer only for what it adds.

Billing period

Free, no account needed

Optional: hosted app at app.cedr.dev

Apache 2.0, no account

Free Forever

Every tool, on your machines.

$0

Forever, on your own plans and keys

  • Route, Council, Guard, Sync and Monitor
  • GitHub Action and CI checks
  • Local MCP server for any agent
  • Apache 2.0, with no licence check

Pro

Your history and routes on every machine.

$6/ dev / month

$60/ dev / year

1 developer minimum

  • Saved routes on every machine
  • 1 year of Monitor and Route history
  • 100 repos on the Sync page
  • 25 API keys
Start Free Trial

Max

Shared rules and rollups for a team.

$19/ dev / month

$190/ dev / year

3 developer minimum

  • Org rules for every repo
  • Team rollups, budgets and spend alerts
  • Shared routes
  • Unlimited members
  • 1,000 repos on the Sync page
Start Free Trial

Ultra

Everything in Max, plus control.

$39/ dev / month

$390/ dev / year

10 developer minimum

  • Everything in Max
  • SAML SSO and audit export
  • 3 years of history
  • 50 API keys
Start Free Trial

Tool by tool

What you get free, and what the app adds.

Route

Free Forever, in the CLI

  • Picks balanced by plan room
  • Verdicts and stats on each machine
  • Local MCP tools

Optional, in the app

  • Activity across machines
  • Saved routes on every machine (Pro)
  • Shared routes for the org (Max)

Council

Free Forever, in the CLI

  • Plan reviewers, one per vendor
  • Merge reports into one gate
  • Runs on your own plans and keys

Optional, in the app

  • Results on the Council page
  • Saved councils (Pro)
  • Hosted MCP merges

Guard

Free Forever, in the CLI

  • Lint, lock and verify
  • CI annotations from the GitHub Action
  • Server change reports

Optional, in the app

  • Every repo on one dashboard
  • Findings and server updates in one list

Sync

Free Forever, in the CLI

  • All of .cedr/, all agents
  • Dry runs and drift checks in CI
  • As many repos as you like

Optional, in the app

  • Drift across up to 1,000 repos
  • Weekly drift digest
  • Org rules for every repo (Max)

Monitor

Free Forever, in the CLI

  • Tokens, spend, sessions, limits
  • Local logs, sent nowhere unless you push
  • Per-model and per-source reports

Optional, in the app

  • History across machines
  • Monthly usage report
  • Team rollups and budgets (Max)

Bench

Free Forever, in the CLI

  • Task suites your tests decide
  • Runs on your own plans and keys
  • A budget on every run
  • Results stay on your machine

Optional, in the app

  • One budget for the organization

US dollars before tax. Your first organization gets a 21-day Max trial, no card needed.

Frequently asked questions

Is the CLI really free?

Yes. Every cedr command line tool is free forever. The app adds shared history, org rules, team rollups and budgets.

Which coding agents does cedr support?

Claude Code, Codex, Antigravity, Muse Code, T3 Code, opencode, Gemini CLI, Copilot CLI, Cursor. Each tool lists the agents it works with. Support for Grok Build, Junie, Kiro and Devin is coming soon.

What does Monitor send to the app?

Counts only: tokens, sessions and plan readings, hourly. Never prompts, code or file paths.

Can Route use API keys as well as plans?

Yes. Turn each provider on as Plan, API key or Both. Route spends plan room first and only then touches a key.

How does the trial work?

Your first organization gets a 21-day Max trial with no card needed. Prices are per developer, in US dollars before tax.

Get your agents out of the woods

No more guessing which agent has room, who checked the work or what it all cost. Start with the free CLI, or try Max for 21 days with no card.