ContextECF CodeLedger Living Demo
A synthetic engineering team. Real CodeLedger output. Ten months of history.
You’re looking at a demonstration of what CodeLedger catches in a real engineering team’s daily work. Sara, Marcus, and Priya are bot personas who open PRs every weekday. Every score, trend, and incident below comes from actual product output running against the acme-platform synthetic monorepo.
Enterprise-class public repo experiments
Validated beyond the demo repo
We run CodeLedger against large public repositories so the demo is not just a synthetic story. The useful result is not perfection. It is evidence: where CodeLedger catches real review risk, where it stays quiet, and where the product needs a sharper signal.
Next.js + React
11 merged PRs, scored blind
5 useful catches, 2 correct silences, 0 hallucinations
Separated systemic signal gaps from repo-local noise and produced concrete fixes.
PostgreSQL
C codebase, revert-heavy sample
Pipeline ran cleanly: no crashes, no indeterminate output, stable JSON
Exposed where JS/TS-shaped risk patterns needed language packs for enterprise C repos.
Postgres UX study
7 real tasks against a 5,000+ file repo
Found a cold-start retrieval gap before claiming victory
That failure mode directly drove selector and prompt-coach hardening.
These are measurement runs, not marketing benchmarks. Selection is pre-registered and filter-based; failures are kept in the story because they improve the product.
See the trust layersWhat problem are we solving?
The Problem — AI coding agents waste 40–60% of their context window on irrelevant files. Every session starts cold. Institutional knowledge lives in people’s heads and disappears when they leave. There is no risk signal before a merge.
The Solution — CodeLedger is a deterministic context control plane for software development. It scores every file in a repository, selects only what the current task requires, captures outcomes, and promotes successful patterns into reusable institutional memory.
The Intelligence Layer — The Task Intelligence Engine is seeded from day one with a curated ontology of golden patterns — distilled from leading engineering teams at organizations including Google, SAP, and Salesforce. Your own earned patterns layer on top, making the system progressively more tailored to your codebase.
Logs are history. Ledger is intelligence.
Recommended click path
- 1. Team Health — 20 seconds to get oriented
- 2. Time Horizon Analytics — the flagship view
- 3. The Auth Incident — the scenario that sells the product
- 4. Fleet Insights — cross-repo view (Enterprise tier)
- 5. Trust Layers — context, audit, and signal health in one proof view
CodeLedger · Intelligent Context AI Inc. · codeledger.dev