Pricing
Know what your AI coding tool actually did.
CodeLedger gives developers better context, engineering leads visibility into AI risk, and CTOs signed proof of what the AI coding system did. Start free. Add your team when ready.
What is CodeLedger?
CodeLedger is the accountability layer for AI coding tools. It tracks which files your AI agent read, what it changed, and whether there is evidence it got things right. The result: your team ships AI-assisted code with confidence instead of guessing whether it worked.
For developers
The CLI picks the right files for any task — no manual hunting, no wasted context.
For engineering leads
See which AI agents are adding risk, which are saving time, and where the codebase is drifting.
For CTOs and security teams
Get a signed certificate proving what the AI coding system did — and what evidence exists for it.
New — Truth Audit Single Run
Not ready for a subscription? Run one audit first.
Pay for a single Truth Audit run. You get access to the MCP server, run generate_truth_report once, and receive a signed certificate you can share with your team or stakeholders. No subscription required.
- ✓ MCP server access — works with Claude Desktop, Cursor, Windsurf
- ✓ One full Truth Audit run on your repo
- ✓ Signed certificate delivered by email
Contributor
Try it on your own code. Free — we just ask for a monthly note back.
- ✓CLI: install, scan your repo, get a ranked file list for any task
- ✓Deterministic context bundles — no AI guessing which files matter
- ✓Session summaries showing recall, precision, and tokens saved
- ✓PR review: catches your highest-severity risks (P0) on every pull request
- ✓GitHub Action setup — one file, works in any repo
- ✓Up to 5 repositories
- ✓Free for Contributors — send one piece of feedback or a solution/improvement idea a month
Your signal history is already building
From your first session, CodeLedger records merge history, drift patterns, and architecture signals. Upgrade to see what it has learned about your codebase.
- 🔒Merge Memory ledgerUnlocks at Team
- 🔒Shared Context LedgerUnlocks at Team
- 🔒Architecture drift snapshotsUnlocks at Team
- 🔒Provenance GraphUnlocks at Enterprise
Team
For engineering teams that need to know what their AI agent actually did.
30-day free trial — no credit card required
- ✓Everything in Contributor
- ✓Full PR review: all severity levels (P0–P2) with risk, drift, and evidence-gap signals
- ✓Truth Audit: 5 signed certificates per 30 days — shareable proof of what the AI did
- ✓Architecture Health Dashboard: see which parts of your codebase are getting riskier over time
- ✓Engineering Intelligence Dashboard: per-agent metrics, hours saved, destabilization trends
- ✓Unlimited seats and repositories
3 ledgers unlocked — 2 still accumulating
Your Provenance Graph and Policy Framework continue building until Enterprise.
- ✓Merge Memory ledger
- ✓Shared Context Ledger
- ✓Architecture drift snapshots
- 🔒Provenance GraphUnlocks at Enterprise
- 🔒Policy Domain FrameworkUnlocks at Enterprise
Enterprise
One org license for your entire engineering team. No per-seat counting.
- ✓Everything in Team
- ✓One shared org token — share with your whole team, no per-seat setup
- ✓Truth Audit certificate program: custom volume, white-label reports
- ✓Unlimited ledger history — no 90-day retention cap
- ✓Custom contract, PO cycles, and SLA
All 5 ledgers fully unlocked
Complete audit-grade provenance and policy governance — ready for board reporting or acquisition review.
- ✓Merge Memory ledger
- ✓Shared Context Ledger
- ✓Architecture drift snapshots
- ✓Provenance Graph
- ✓Policy Domain Framework
Prompt Coach · Signal Architecture
The advice gets better as your team uses it more
Before your AI agent starts on a task, the Prompt Coach checks whether the request is clear enough to succeed — catching missing context, scope contradictions, and risky patterns upfront. On Contributor it uses a curated industry baseline. On Team it learns from your repo's own history. The quality of that coaching improves at every tier.
Good — calibrated from published AI research
A signal library distilled from Anthropic, OpenAI, Google, AWS Bedrock, GitHub Copilot, and SWE-bench. Ships on day one, no configuration needed.
- ✓ 7 signal categories: meta, safety, chain-of-thought, few-shot, structured output, self-critique, governance
- ✓ Task archetypes: auth, migrations, refactors, dependency upgrades, APIs
- ✓ Evidence-backed signal weights from published research
Better — tuned to your codebase
The public baseline plus patterns learned from your team's actual shipping history. Signal weights calibrate automatically — the more you use it, the more accurate it gets.
- ✓ Everything in Contributor
- ✓ Team-learned signal promotions
- ✓ Weights auto-tuned to your repo's task history and production outcomes
- ✓ Shared pack synced across all team members automatically
Best — network-calibrated + domain-specific
Org intelligence plus opt-in cross-org signals from the broader CodeLedger network. Custom domain packs for fintech, healthcare, and infrastructure.
- ✓ Everything in Team
- ✓ Anonymized cross-org pattern signals (opt-in)
- ✓ Enterprise governance signal layer (compliance, auditability, access control)
- ✓ Private pack mode — signals never leave your VPC
The Prompt Coach starts strong on day one. At Team it learns your codebase. At Enterprise it learns from the network — so every task you describe gets better advice than the one before it.
Everything included at each tier
| Feature | Contributor | Team | Enterprise |
|---|---|---|---|
| Context & Bundles | |||
| Deterministic file scoring | ✓ | ✓ | ✓ |
| Context bundles & activation | ✓ | ✓ | ✓ |
| Shadow file expansion | ✓ | ✓ | ✓ |
| Session summaries & recall metrics | ✓ | ✓ | ✓ |
| Ledgers & Signal Architecture | |||
| Merge Memory ledger | Building | ✓ | ✓ |
| Shared Context Ledger | Building | ✓ | ✓ |
| Architecture drift snapshots | Building | ✓ | ✓ |
| Cross-session memory | Building | ✓ | ✓ |
| Provenance Graph | Building | Building | ✓ |
| Policy Domain Framework | Building | Building | ✓ |
| Collective Intelligence | |||
| Prompt Coach (public signal baseline) | ✓ | ✓ | ✓ |
| Team-learned signal promotions | — | ✓ | ✓ |
| Outcome-tuned signal weights | — | ✓ | ✓ |
| Deploy outcome feedback loop | — | ✓ | ✓ |
| Anonymized cross-org signals (opt-in) | — | — | ✓ |
| Custom domain signal packs | — | — | ✓ |
| Verification & Intelligence | |||
| PR Review Intelligence (P0 only) | ✓ | ✓ | ✓ |
| Full Review Intelligence (P0–P2) | — | ✓ | ✓ |
| PR risk / drift / evidence-gap signals | — | ✓ | ✓ |
| Architecture Health Dashboard | — | ✓ | ✓ |
| Engineering Intelligence Dashboard | — | ✓ | ✓ |
| Local Truth Audit report | — | ✓ | ✓ |
| Signed external Truth Audit certificates | — | 5 / 30 days | Custom volume |
| Truth Audit billing/account webhook | — | Metered | Private endpoint |
| Governance | |||
| Audit exports (JSON + Markdown) | — | — | ✓ |
| SSO / SAML (roadmap) | — | — | Planned |
| On-prem / VPC deployment | — | — | ✓ |
| Unlimited ledger retention | — | — | ✓ |
| Support | |||
| Community support | ✓ | ✓ | ✓ |
| Email support | — | ✓ | ✓ |
| Dedicated Slack channel + SLA | — | — | ✓ |
| Custom MSA & PO cycles | — | — | ✓ |
Common questions
Still have questions? Talk to us.
We'll walk you through how CodeLedger works with your stack, answer compliance questions, and scope a trial for your team.
Contact sales →