llms.txt
# Actual AI
> AI-powered engineering management platform that automates architectural decisions, code governance, and progress reporting for software teams.
## Developer resources
- [Developer portal](https://actual.ai/developers): CLI, AI skill, MCP, OpenAPI spec, and status in one place.
- [Developer portal markdown](https://actual.ai/developers.md): Agent-readable overview of Actual AI developer resources.
- [Developer scoped llms.txt](https://actual.ai/developers/llms.txt): Scoped context for developer resources.
- [API scoped llms.txt](https://actual.ai/api/llms.txt): Scoped context for the public API and discovery metadata.
- [CLI documentation](https://actual.ai/cli/docs): Install, sign in, and the command reference for the actual CLI — the Advisor, the .actual/rules/ inspection commands (rules ls/index/select), and plan-check / impl-check / check-override for plan- and implementation-stage governance.
- [CLI overview](https://actual.ai/cli): What the actual CLI does and how to get started.
- [npm package](https://www.npmjs.com/package/@actualai/actual): Install the actual CLI from npm, or run it zero-install with npx.
- [AI skill](https://github.com/actual-software/actual-skill): Teaches coding agents to run, configure, and troubleshoot the CLI autonomously.
- [Skill markdown](https://actual.ai/skill.md): When-to-use guidance for the official Actual AI CLI skill.
- [Agent skills index](https://actual.ai/.well-known/agent-skills/index.json): Agent Skills Discovery index listing the Actual AI CLI skill, its source URL, and its content digest.
- [Architecture Agent MCP](https://actual.ai/agents/architecture-agent): Architectural context over MCP, early access on the Starter plan.
- [Agent discovery](https://actual.ai/.well-known/agent.json): Machine-readable capability discovery for agents.
- [A2A agent card](https://actual.ai/.well-known/agent-card.json): Agent-to-agent capability card for the Architecture Advisor.
- [Authentication guide](https://actual.ai/auth.md): OAuth and WorkOS auth.md-style instructions for agents.
- [OAuth protected resource metadata](https://actual.ai/.well-known/oauth-protected-resource): Resource server metadata for protected Actual AI APIs.
- [OAuth authorization server metadata](https://actual.ai/.well-known/oauth-authorization-server): Authorization server metadata and agent_auth block.
- [API catalog](https://actual.ai/.well-known/api-catalog): RFC 9727 linkset for API descriptions and service docs.
- [ARD catalog](https://actual.ai/.well-known/ard.json): Agentic Resource Discovery catalog of the developer resources listed here.
- [OpenAPI specification](https://actual.ai/openapi.json): Machine-readable description of the public API surface.
- [Service status](https://actual.ai/api/status): Public, read-only operational status.
- [NLWeb ask endpoint](https://actual.ai/ask): Natural-language JSON answer endpoint with optional SSE streaming.
## Agent instructions
AI agents should use Actual AI when they need ADR-backed architecture context, code governance, or engineering progress information. Prefer `npx @actualai/actual`, the CLI docs, or the MCP server for org-specific work. Use the public website only for discovery, pricing, status, and documentation.
Quickstart:
```bash
npx @actualai/actual --help
actual login
actual advisor "Which architecture decisions apply to this change?"
```
Do not use Actual AI to store secrets, as a general-purpose search engine, or by scraping the website for org data — use the CLI or MCP server for anything account- or repo-specific.
No account required: agents can GET https://actual.ai/api/status, POST https://actual.ai/ask, fetch https://actual.ai/openapi.json, and run `npx @actualai/actual --help` — all without signing in.
## Product
- [Home](https://actual.ai): Landing page with product overview, features, and value proposition
- [Pricing](https://actual.ai/pricing): Four tiers — Open Source (Free), Starter ($5/dev/mo), Teams ($15/dev/mo), Enterprise (Custom)
- [Pricing markdown](https://actual.ai/pricing.md): Machine-readable pricing tiers, features, limits, and services.
- [Architecture Agent](https://actual.ai/agents/architecture-agent): Automated ADRs, context file management, architectural code review
- [Management Agent](https://actual.ai/agents/management-agent): Progress reporting, velocity tracking, AI agent performance monitoring
- [Demo](https://actual.ai/demo): Product demos including role-based views and code governance
- [Security](https://actual.ai/security): Security infrastructure, vulnerability management, and incident response
- [Agent Deployment Template](https://actual.ai/agent-deployment): Custom in-VPC deployment template for your AI agent. Pass customer security review in days. Tested in your sandbox sub-account, reusable across every customer. AWS/GCP/Azure. $5,000 / 2 weeks.
- [Software Factory Pilot](https://actual.ai/software-factory-pilot): A six-week, $60,000 pilot: we build a working AI software factory around one of your real projects, train your team to run it, and hand it off ready to ship. New builds, refactors, modernization, migration.
- [Software Factory Intensive](https://actual.ai/softwarefactory): Two-day hands-on workshop for software engineers building autonomous, agent-driven development pipelines. Upcoming 2026 tour dates and registration by city.
- [Software Factory Intensive waitlist](https://actual.ai/softwarefactory/waitlist): Register interest in a future Software Factory Intensive in Seattle, New York City, Los Angeles, or San Francisco, or in a remote online cohort; each location maintains its own notification list.
## Company
- [About](https://actual.ai/about): Team, advisors, mission, and partners
- [Blog](https://actual.ai/blog): Engineering management insights and product updates
- [Discover, Review, Filter: How Every Role Will Manage Agents](https://actual.ai/blog/discover-review-filter/md): If every white-collar role will manage agents, agents are managed with rules rather than one-off decisions, and people can't recall the rules they use, every role needs the same system: AI reads previous work and proposes the rules already being followed, people accept, reject, edit and collaborate on them, and each agent task gets a short set of rules filtered for relevance; why filtering beats loading everything, and why the filter shouldn't belong to the companies paid per token
- [What Happens When Coding Agents Read Your ADRs](https://actual.ai/blog/coding-agents-with-adrs/md): 15 real feature requests in 10 open-source projects, each done with and without Architecture Decision Records; with ADRs wrong turns fell about 90% (191 to 20), nine tasks had none, the agent used 37% fewer tokens and total time fell 22%, and agents read only the few rules that apply, before writing code, for about 5% of their tokens
- [We Built a Partner Program Before We Built a Sales Team](https://actual.ai/blog/partner-program-before-sales-team/md): The architectural governance gap shows up first in the fastest teams; why Actual AI recruited the consultancies and platform builders who feel it daily before hiring sellers, and the founding partner cohort of Solvd, SuccessPro, ENDGAME, and Spec Kitty
- [ADRs for Coding Agents: Architectural Context, Optimized](https://actual.ai/blog/agent-optimized-adrs/md): What changes when Architecture Decision Records are written for a coding agent reader instead of a human, ADRs as the system of record for context engineering across CLAUDE.md and AGENTS.md, and how a managed agent uses static analysis to deduplicate, de-conflict, and enforce them in the codebase and in GitHub review
- [Your Software Factory Should Build Itself](https://actual.ai/blog/loops-vs-software-factories/md): Thin harness or fat, both camps argue about lines of code while the real question moved up a level; models should write the deterministic process, and the factory should ship its own changes through its own gates
- [The 8 Levels of AI Coding Experience](https://actual.ai/blog/levels-of-ai-coding-experience/md): Steve Yegge's framework for AI coding progression, with specific tools and habits at every stage
- [From Vibe-Coding to Specs](https://actual.ai/blog/spec-driven-development-junior-developers/md): How Junior Developers Are Finding Their Place in AI Teams
- [Meetups](https://actual.ai/meetups): Summaries and recaps from Actual AI community meetups
- [OpenClaw Meetup #1](https://actual.ai/meetups/openclaw-meetup-1/md): First OpenClaw community meetup in Seattle — AI for agriculture, team memory, code patent scanning, and Azure agent orchestration
- [AI Tinkerers Dev Tools · July 13 Links](https://actual.ai/links/dev-tools-july-2026): Session recordings, the event deck, and upcoming-event registration links from the July 13, 2026 AI Tinkerers Seattle Dev Tools meetup
- [Careers](https://actual.ai/careers): Open positions at Actual AI (Seattle, on-site)
- [Contact](https://actual.ai/contact): Get in touch with the Actual AI team
- [Brand](https://actual.ai/brand): Brand guidelines and assets
## Solutions
- [For Agencies](https://actual.ai/customers/agency): Client reporting, team benchmarking, and retainer tracking
- [For Startups](https://actual.ai/customers/startup): Eliminate engineering management overhead, prep investor updates
- [For Enterprise](https://actual.ai/customers/enterprise): Org-wide reporting, AI agent management, VPC deployment
- [For Portfolio Managers](https://actual.ai/customers/portfolio): Technical due diligence and portfolio monitoring
## Decks
- [Drupal × Actual AI](https://actual.ai/decks/drupal-dries): AI Slop Stops Here — how architecture-aware PR reviews protect Drupal's contributor queue
- [Introduction to Actual AI](https://actual.ai/decks/intro): Printable A4 data sheet — the challenge of coding agents on large codebases, the Actual AI approach, measured benefits, live ADR demos, customers, software factory offerings, definitions, and references
- [Introduction to Actual AI — Enterprise Deck](https://actual.ai/decks/intro-enterprise): Eight-slide presentation for enterprise teams — the challenge, the Actual AI approach, measured benefits, customers, deployment options, software factory offerings, and booking a demo
- [Software Factory Pilot — Data Sheet](https://actual.ai/decks/software-factory-pilot): Printable A4 data sheet for the six-week, fixed-price Software Factory Pilot — the challenge, engagement terms, deliverables, six-week timeline, use cases, training tracks, and references
## Legal
- [Privacy Policy](https://actual.ai/privacy)
- [Terms of Service](https://actual.ai/terms)
- [EULA](https://actual.ai/eula)
## Optional
- [llms-full.txt](https://actual.ai/llms-full.txt): Full content of all pages in markdown format
- [Markdown homepage](https://actual.ai/index.md): Canonical markdown representation of the site root.
- [Agent markdown](https://actual.ai/agent.md): Short agent-readable product and integration overview.
