01
Understand
Living Twin
Repositories, entities, and relationships loaded from the twin
Physical code, conceptual memory, commits, releases, ADRs, sessions, dependencies, and impact paths in one graph.
Enterprise AI SDLC control plane
Built with IBM Bob
Living code digital twin
Govern every software change from idea to evidence.
BobFactory turns repositories, architecture decisions, releases, commits, agent work, and Bob evidence into one operating system for AI-native software delivery.
Live control path
Select any repository entity, then explain, simulate, execute, and prove.
The page is driven by the current digital twin, so the same workflow applies to enterprise, startup, platform, and open-source repositories after ingestion.
RationaleAsk Bob explains purpose from code, ADRs, commits, and relationships.
ImpactThe impact engine traces direct and transitive consumers across repositories.
ProofAudit keeps graph, chat, impact, agent, and Bob evidence reviewable.
Live
Repos in governed twin
Live
Twin entities
Live
Relationships
Live
Evidence files
The control loop
One enterprise workflow, five governed stages
Strong competitors explain code. BobFactory shows the whole operating loop: memory, prediction, execution, and proof.
02
Explain
Ask Bob
Ask why any selected entity exists
Answers include rationale, intent, commit history, ADR governance, direct callers, and dependent repositories.
03
Predict
Impact Engine
Blast radius across all ingested repositories
Rename simulations produce blast radius, breaking-change warnings, migration steps, and regression scope.
04
Execute
Agent Factory
Plan + code + tests + review package
Planner, coder, tester, and reviewer agents create a non-mutating execution package with provenance hashes.
05
Prove
Evidence Ledger
26 Bob evidence files surfaced
Bob session exports, internal monologue, audit events, graph snapshots, ADRs, and generated artifacts remain reviewable.
Judge-visible proof path
The product tells one complete story from live repository data.
Ask whySelect an entity and ask Bob to explain why it exists.
Simulate changeRun modify, delete, or rename impact to surface consumers and migration steps.
Travel timeOpen a snapshot or release to reconstruct historical twin state.
Execute with agentsThe orchestrator emits a plan, generated code, generated tests, review, validation, and provenance hashes.
Prove itAudit shows Bob sessions, internal monologue, graph snapshots, chat, impact, refactor, and orchestration events.
Scoring alignment
Built with Bob workflows, Bob reports, internal monologue, and evidence surfaced in-product.
Cuts onboarding, release risk, legacy migration risk, audit drag, and AI governance gaps.
Combines code digital twin, historical memory, impact simulation, and execution provenance.
The control loop is visible from any ingested repository.
Enterprise value
Why this wins inside real teams
Onboarding
New engineers ask the twin for history, ownership, and current role before touching code.
Release safety
Every proposed change gets downstream dependency, breaking-change, and migration evidence.
AI governance
Generated plans, tests, reviews, sessions, and audit entries are attached to software memory.
Modernization
Legacy APIs become managed migration programs instead of invisible tribal knowledge.