Anvil
The self-hosted system of record for organizations, projects, Product Engineers, memory, journals, credentials by reference, Spark trust, licensing, backups, and administration.
The Engineering Operating System
Ephemeral sessions. Permanent engineers.
Forge turns replaceable AI reasoning sessions into permanent Product Engineers with verified knowledge, secure execution, and traceable engineering history.
Your AI session should be disposable. Your engineering organization should not be.
Working Memory
Verified facts only
Decisions, commits, tasks, artifacts, reviews, and operational evidence rebuild every session.
Conversation is the interface. Engineering knowledge is the product.
Why Forge Exists
Your engineers should not forget the product every time a conversation ends.
Generic AI chat remembers a conversation. Forge remembers the engineering organization.
A cleared conversation, provider change, model upgrade, or exhausted context window should never erase product ownership, architectural decisions, validated procedures, or operational evidence.
Engineering Chain of Provenance
Know why every decision was made.
Forge preserves evidence of what engineering work actually occurred, without exposing private chain-of-thought.
Director directive
Product Engineer session
Spark execution
Command, test, or deployment result
Artifact hash
Review or approval
Verified Engineering Intelligence
Forge Product Family
One operating system. Clear responsibilities.
Hammer
The Engineering Director's workbench for checking out Product Engineers, assigning missions, directing work conversationally, reviewing outcomes, and reconnecting to multiple Anvils.
Spark
The secure execution runtime installed where work lives. Spark has no AI and owns no knowledge; it executes authorized terminal, Git, build, test, deploy, and artifact operations.
Stewart
The operational guide for onboarding, Spark enrollment, Hammer connection, backups, health questions, and verified Forge support knowledge.
Sentinel
An optional Infrastructure Reliability Engineer for heartbeat health, operational state changes, incident evidence, reliability knowledge, and structured root-cause investigations.
Onboard the engineer, not the chat history.
Forge interviews existing AI assistants and captures durable truth: product ownership, architecture, business rules, repositories, environments, risks, validation contracts, and protected boundaries.
Change the reasoning engine, not the engineer.
Use OpenAI, Anthropic, Codex, or future providers without surrendering the Product Engineer's identity, ownership, memory, or institutional knowledge.
Preserve evidence, not just summaries.
Engineering Artifacts keep inventories, logs, build output, deployment evidence, screenshots, and structured results with provenance, hashes, and verification records.
Who Forge Is For
Teams where engineering knowledge is too valuable to live in scattered memories.
Engineering Directors and CTOs managing multiple products, repositories, infrastructure environments, and review streams.
Software consultancies and MSPs that need customer isolation, billable sessions, repeatable onboarding, and auditable engineering history.
Internal IT and platform teams modernizing legacy systems, reports, databases, manual processes, and undocumented server configurations.
Security-conscious organizations that want AI-assisted engineering while keeping permanent knowledge and execution control inside customer-managed infrastructure.
Generic AI Tools
Temporary task execution.
- Remember a conversation.
- Own a temporary task.
- Run commands through an agent layer.
- Leave project truth scattered across tools.
Forge
Permanent engineering organization.
- Create Product Engineers with enduring ownership.
- Rebuild sessions from verified Working Memory.
- Connect direction, execution, evidence, reviews, and decisions.
- Preserve institutional knowledge that belongs to the organization.
Forge
Stop managing prompts. Start leading engineers.
Forge allows Engineering Directors to lead AI Product Engineers the same way they lead senior human engineers: with direct conversation, clear ownership, verifiable evidence, and accountable outcomes.