ForgeLayer

AI software control layer

AI Software Teams. Controlled.

ForgeLayer helps teams review, route, record, and control AI-generated software work before it reaches production.

MergeGuard, RepoBrain, AI Work Ledger foundations, PromptForge v1, and AgentOS planning modules give AI software teams a safer control plane without live agent execution.

Explore Private BetaOpen WorkspaceExplore MergeGuardExplore RepoBrainExplore PromptForgeExplore AgentOSExplore LedgerView Review RunsExplore Control FlowExplore AnalyticsGenerate a reviewTry a demo scenarioStart private betaOpen demo scriptCheck readinessView Trust CenterRepository Connection
MergeGuard activePrivate betaLedger localRepoBrain policyGitHub posting disabledBilling not live

Account

Authentication disabled

Public production authentication remains fail closed. Deterministic demo access is separate and does not create a trusted identity.

Working foundations

Control AI code work from prompt to review run

Prompt -> Diff -> Repo policy -> Risk report -> PR comment -> Work record.

01

MergeGuard

Reviews diffs and review runs for risk, missing proof, and merge readiness.

02

RepoBrain

Applies repo policy, protected paths, sensitive paths, and review requirements.

03

AI Work Ledger

Starts the system of record with local records and stored review-run summaries.

04

PromptForge v1

Scores prompts, suggests rewrites, and generates deterministic security tests.

05

AgentOS Foundation

Provides registry, routing hints, and deterministic planning/control modules.

06

Task-to-Agent Workflow

Creates draft routing plans, approval checks, blocked actions, and next steps.

07

Agent Work Records

Documents draft agent work, evidence, risks, approvals, and merge readiness.

08

Agent Comparison

Compares seeded agents by fit, safety, autonomy, review need, and control risk.

09

Review Runs

Displays stored safe GitHub review summaries with filters and detail views.

Try a demo scenario

Load a reviewer-ready example

Pick one of four stable private-beta scenarios, then generate the report.

Demo flow

Show the control loop in five minutes

Beta demo guidePricing preview
  1. Load docs, database, auth, or prompt-injection work.
  2. Load example repo rules for RepoBrain policy context.
  3. Generate risk, merge recommendation, and follow-ups.
  4. Copy the PR comment and show the AI Work Record.

Testing ForgeLayer?

Send feedback after the walkthrough.

Open feedback template

Control plane input

Generate ForgeLayer Report

Paste a task and diff to create a merge decision, PR comment, work record, and repo-context check.

Private beta mode: billing is not live, GitHub posting is disabled unless enabled server-side, and known Supabase/repo-rules fetch issues do not affect this manual demo.

Public PR .diff URL

Paste a GitHub PR .diff URL or paste the diff directly above.

Only public GitHub PR .diff URLs are supported. Private repos require a future GitHub App.

Repo rules / forgelayer.yml

Repo rules make ForgeLayer stricter for company-specific rules like auth tests, migration rollback plans, and config approval.

No rules provided0 rule lines detected

RepoBrain v2

No repo rules loaded yet

No rules

No RepoBrain profile yet. Paste forgelayer.yml or repo rules to derive repository policy context.

Source: noneCategories: none

Report output

Start with a demo scenario, pasted diff, or public PR .diff URL.

ForgeLayer reviews pasted or imported diff text only. It will create a merge decision, AI Work Ledger record, RepoBrain context summary, and PR comment preview.

AI Work Ledger

Recent work records

Ledger v1 is stored only in this browser using localStorage. It is not synced, uploaded, or ready for enterprise audit storage yet.

No AI Work Ledger records yet. Generate a report to save the first local work record in this browser.

Latest 20 work records are kept in this browser.