# Amos

> Amos — an AI workforce platform from GDI Labs.

## Docs

- [Multi-agent workforce](https://docs.startamos.io/capabilities/ai-workforce.md): Amos runs as a team, not a single agent. Each request is classified, routed to the right specialist, executed, and verified — all without you needing to manage the choreography.
- [Audit-ready by default](https://docs.startamos.io/capabilities/audit-and-cost.md): Amos was designed for places where you need to explain what an AI did and why. Every action is recorded.
- [AWS Pulse Report](https://docs.startamos.io/capabilities/aws-pulse-report.md): A daily-cadence operations scenario that audits an AWS account via read-only API calls, synthesizes a Markdown summary, and posts a Block Kit version to Slack — end-to-end, with no human in the loop.
- [Declarative agent flows](https://docs.startamos.io/capabilities/declarative-agent-flows.md): Amos runs every paid job through a scenario: a versioned YAML file that describes the agent flow as a DAG of levels. Adding a new flow is a PR that touches knowledge-hub/scenarios/; no worker deploy is required once the underlying primit...
- [Frontend Authentication & Authorization](https://docs.startamos.io/capabilities/frontend-authentication.md): ## What it enables
- [Hide Workflow](https://docs.startamos.io/capabilities/hide-workflow.md): ## What it enables
- [Grounded in your knowledge](https://docs.startamos.io/capabilities/knowledge-grounding.md): Every Amos answer is anchored to your organization's documents, not the model's generic training data. Cite-and-source is built in.
- [Free and paid models, together](https://docs.startamos.io/capabilities/multi-model.md): Amos doesn't ask you to pick a model upfront. It picks per step, balancing cost against the difficulty of the work.
- [Project Console (/projects)](https://docs.startamos.io/capabilities/project-console.md): ## What it enables
- [Requirements-Clarification Gate](https://docs.startamos.io/capabilities/requirements-clarification-gate.md): ## What it enables
- [Scenario-Aware Job Output Panel](https://docs.startamos.io/capabilities/scenario-aware-job-panel.md): The workflow detail panel renders model output per scenario. Quick-QA jobs no longer dump their JSON envelope as raw text under a "Project output" header — the answer renders as a card with a confidence pill and a sources list. Markdown...
- [Solutions Catalog & Template-Last Bootstrapping](https://docs.startamos.io/capabilities/solutions-catalog.md): ## What it enables
- [Universal Task Input](https://docs.startamos.io/capabilities/universal-input.md): ## What it enables
- [Universal User-Feedback System](https://docs.startamos.io/capabilities/user-feedback-system.md): ## What it enables
- [Agent Workbench panel](https://docs.startamos.io/developers/components/agent-workbench.md): The left-side workflow list in the Amos dashboard (apps/frontend/components/workflow/panels/jobs-list-panel.tsx). It is the entry point for inspecting any running, queued, or recently-completed workflow.
- [External-source connectors](https://docs.startamos.io/developers/components/connectors.md): Three CronJobs pull content from sources that live outside the curated hub repo and write it into the same retrieval substrate. Notion gives Amos access to product docs. Slack gives it conversation context. GitHub gives it raw source cod...
- [Per-agent diaries](https://docs.startamos.io/developers/components/diaries.md): Each worker agent writes back to the knowledge hub. A diary entry is a small markdown document the orchestrator emits after every task — and after every task that recovered from a non-trivial error — so the agent's experience accumulates...
- [External Tools Platform](https://docs.startamos.io/developers/components/external-tools-platform.md): A registry + dispatch layer that lets scenarios call out to remote services (AWS, Slack, future REST connectors) alongside the existing workspace file tools. Lives at worker/worker/runtime/tools/.
- [Frontend](https://docs.startamos.io/developers/components/frontend.md): Next.js 16 / React 19 dashboard. Hosted on Vercel.
- [Ingest](https://docs.startamos.io/developers/components/ingest.md): The pipeline that pushes the knowledge hub into Qdrant. Runs locally for one-off pushes and in-cluster as a scheduled Job.
- [Intent Classifier](https://docs.startamos.io/developers/components/intent-classifier.md): The intent_classifier is a meta scenario that routes free-text prompts to the right downstream scenario. It is the first user of the scenario chaining primitive in the orchestrator — when the classifier picks a target, the orchestrator r...
- [Job Telemetry Snapshot](https://docs.startamos.io/developers/components/job-telemetry.md): ## What it covers
- [L4 Tool Loop — ask_user](https://docs.startamos.io/developers/components/l4-ask-user.md): The L4 agentic tool loop (ToolLoopLevelRunner in worker/worker/graph/scenarios/runners.py) can now pause itself mid-loop to ask the human a clarifying question. Before this change the loop only had file-mutation tools and finish; on unde...
- [Knowledge Hub MCP](https://docs.startamos.io/developers/components/mcp-amos-kb.md): A read-only MCP server that exposes the Amos knowledge hub to any MCP client — Claude Desktop, Claude Code, partner tools, agents — through a single uniform contract. The same image serves three transports; choose at runtime.
- [Mother AI](https://docs.startamos.io/developers/components/mother-ai.md): The ingest service. Rust + Axum, stateless beyond Postgres-backed projects and workflows.
- [Notion ingest — v3 (authors, comments, databases, backfill)](https://docs.startamos.io/developers/components/notion-ingest.md): v3 lifts the Notion connector from "we have page bodies" to "we have the whole corpus with author lineage, threaded discussions, database rows, and a cross-link graph." The v1/v2 mechanics (CronJob, search-by-last_edited_time cursor, 3 r...
- [Scenarios Runtime](https://docs.startamos.io/developers/components/scenarios-runtime.md): The scenarios runtime is the worker's pluggable engine for orchestrating multi-step agent flows. It replaced the hardcoded L1→L2→L4→L3 LangGraph subgraph (deleted in [worker/worker/graph/agents/hierarchy.py][hierarchy-removed] in May 202...
- [Worker](https://docs.startamos.io/developers/components/worker.md): The orchestrator. A Python LangGraph service that consumes the job queue, augments prompts with retrieved context, runs every paid job through the scenario runtime, dispatches to free or paid models, and streams typed events back through...
- [System overview](https://docs.startamos.io/developers/overview.md): Amos is a queue-first AI orchestration platform. It runs a multi-agent hierarchy against free and paid models, augmented by a retrieval-grounded knowledge hub, with full audit trails and per-project workspace provisioning.
- [The agent hierarchy](https://docs.startamos.io/how-it-works/agent-hierarchy.md): > Implementation note (2026-05-13). The L1/L2/L3/L4 split below > remains the conceptual shape of code-build runs but is no longer > hardcoded in the worker. Every flow — code, writing, research, > security, Q&A — is now a versioned YAML...
- [The knowledge hub](https://docs.startamos.io/how-it-works/knowledge-hub.md): The knowledge hub is the organizational memory Amos's agents draw on. Every answer is grounded here.
- [Project Identity — Display Name vs Slug](https://docs.startamos.io/how-it-works/project-slug-identity.md): Two users can now create projects with the same display name ("test", "Hello world", …) without colliding on GitHub repos, Vercel projects, or local workspace dirs. The data model splits a project into two fields: name (free-text, what u...
- [Scenario execution](https://docs.startamos.io/how-it-works/scenario-execution.md): This document walks a single job through the scenario runtime end-to-end: from the worker receiving a payload to the terminal ScenarioResult and the event stream that lands on the SSE channel. It's the companion to the API reference in S...
- [Amos roadmap](https://docs.startamos.io/roadmap.md): Amos is built as a transition, not as a finished product. Each phase shifts more of the operational center from humans to AI, with humans moving up — to policy, audit, and intervention — rather than out.
- [Welcome to Amos](https://docs.startamos.io/welcome.md): Amos is an AI workforce for your organization. Not one chat assistant — a coordinated team of specialized AI agents that plan, build, review, and deliver work, grounded in your own knowledge.
- [What is Amos](https://docs.startamos.io/what-is-amos.md): Amos is the institution itself, in software form — for companies where AI is the workforce and humans are the governance layer. It is not "enterprise AI tooling," and it is not another way to make humans 10× more productive.

## OpenAPI Specs

- [openapi](https://docs.startamos.io/api-reference/openapi.json)