# UseAgent > A multi-harness agent platform. An event-sourced control plane drives replaceable coding engines in isolated sandboxes and renders every run from one canonical event log. ## Docs - [Overview](https://useagent.org/docs): A multi-harness agent platform. An event-sourced control plane drives replaceable coding engines in isolated sandboxes and renders every run from one canonical event log. ## Getting started - [Getting started](https://useagent.org/docs/getting-started): What the platform is made of, and the shortest path from a clone to a running stack. - [Quickstart](https://useagent.org/docs/getting-started/quickstart): Install the two apps and bring up the backend and frontend dev servers. - [Repository map](https://useagent.org/docs/getting-started/repository-map): Where each part of the platform lives, from the two apps to the shared contract packages. ## Features - [Features](https://useagent.org/docs/features): Why UseAgent. The agent contract, the trust lane, the reliability story, replaceable engines, the workspace, the self-improvement loop, and the product experience, each as a deep dive. - [The agent contract](https://useagent.org/docs/features/agent-contract): The canonical spec. One provider-neutral event grammar, a versioned envelope, a negotiated capability map, and the delivery guarantees every engine must satisfy. - [Trust and control](https://useagent.org/docs/features/trust-and-control): Autonomy with a hard boundary. Capability tokens, a gateway on a restricted database role, one-shot human approvals, write-only encrypted secrets, and org scoping enforced below the application. - [Reliability](https://useagent.org/docs/features/reliability): Work that survives. Runs outlive backend restarts, reconnects replay exactly what live viewers saw, retries are idempotent, and releases pass real engine journeys before they ship. - [Engines](https://useagent.org/docs/features/engines): The inner loop is rented, not built. Codex, Claude, and OpenCode behind one neutral adapter boundary, capability-gated surfaces, and a subscription relay that keeps OAuth out of the sandbox. - [The workspace](https://useagent.org/docs/features/workspace): A real Linux workstation per thread. Repositories, terminal, a visible desktop with browser automation and recording, injected secrets, artifacts, and editable workpieces. - [Self-improvement](https://useagent.org/docs/features/self-improvement): Runs that compound. Agents recall the organization's know-how at the start, propose what they learned at the end, and nothing publishes without a human accepting it. - [Product experience](https://useagent.org/docs/features/product-experience): A native surface owned end to end. One timeline grammar for every engine, a side rail of live panes, a command palette, four themes, a mobile shell, and Slack as a first-class channel. ## Product - [Product](https://useagent.org/docs/product): The app you actually use. Three tiers of surface, from starting a task to watching it run to managing everything an organization accumulates. - [Task composer](https://useagent.org/docs/product/task-composer): Where a run begins. The composer at /agent/new, and the one Send, Steer, Stop control that also drives replies inside a session. - [Session view](https://useagent.org/docs/product/session-view): The screen where a run comes alive. A structured timeline grammar plus a side rail of surfaces for the browser, terminal, files, diffs, agents, editors, and workpieces. - [Dashboard](https://useagent.org/docs/product/dashboard): The workspace hub. Recent runs, activity over time, fleet lanes per project, and where usage stands against limits. - [Skills and playbooks](https://useagent.org/docs/product/skills-and-playbooks): The reusable-procedure catalog. Where an organization's know-how is stored, imported from GitHub, and surfaced by relevance each turn. - [Knowledge and wiki](https://useagent.org/docs/product/knowledge-and-wiki): Reference material an agent can search, and published documents generated from a repository. The /knowledge and /wiki surfaces. - [Memory hub](https://useagent.org/docs/product/memory-hub): Browse what UseAgent remembers, audit what it captured, and see what it recalled. The /memory surface, its scopes, and its delivery safety. - [Learnings review](https://useagent.org/docs/product/learnings): The human gate on self-improvement. Where proposed knowledge drafts and skill revisions wait for an admin to accept or dismiss. - [Artifacts](https://useagent.org/docs/product/artifacts): The files runs produce and the editors that open them. The /artifacts hub, workpiece editing, and revision proposals. - [Automations](https://useagent.org/docs/product/automations): Runs that start on a schedule or a trigger, through the same command lane as a person. The /agent/automations surface. - [Secrets](https://useagent.org/docs/product/secrets): The encrypted organization secrets manager. Values are write-only, encrypted at rest, and injected into every sandbox as environment variables. - [Settings](https://useagent.org/docs/product/settings): Workspace configuration in one place. General settings, usage, the team, and the provider connections that authenticate engines. - [Review and integrations](https://useagent.org/docs/product/review-and-apps): Reviewing pull requests across connected repositories, and the integrations available to the workspace and to agents. - [Themes and design](https://useagent.org/docs/product/themes-and-design): The token-driven design system behind every surface, the four themes, and the places you can explore both. ## Concepts - [Concepts](https://useagent.org/docs/concepts): The handful of ideas the rest of the platform is built on, and where each one is defined in depth. - [Runs, threads, and commands](https://useagent.org/docs/concepts/runs-threads-commands): The units of work. A command starts a run, a run is one turn, and a thread is the conversation that ties runs together. - [Events and streaming](https://useagent.org/docs/concepts/events-and-streaming): Canonical events, native frames, and the two server-sent streams that carry a run to the browser. - [Engines and adapters](https://useagent.org/docs/concepts/engines-and-adapters): How Codex, Claude, and OpenCode are driven behind a neutral adapter, the runtime path versus ACP, and the Codex subscription relay. - [Sandboxes and desktop](https://useagent.org/docs/concepts/sandboxes-and-desktop): The isolated Linux workstation a run executes in, the provider behind it, warm pools, the visible desktop, and how a run survives a backend restart. - [Gateway, tools, and approvals](https://useagent.org/docs/concepts/gateway-tools-and-approvals): The trust boundary that does privileged work for a sandbox, the capability tokens that bound it, the in-run tool families, and the human approval lane. - [Knowledge and learning](https://useagent.org/docs/concepts/knowledge-and-learning): Skills, playbooks, knowledge, wiki, and memory, plus the human-gated loop that turns a successful run into a reusable procedure. - [Artifacts and workpieces](https://useagent.org/docs/concepts/artifacts-and-workpieces): The files a run produces, the four editable workpiece kinds, uploads coming the other way, and the outbox that delivers follow-up work safely. ## Architecture - [Architecture](https://useagent.org/docs/architecture): The settled architecture and why each property is load-bearing. - [Request flow](https://useagent.org/docs/architecture/request-flow): One turn, end to end. From an accepted command to a rendered timeline, and the promote-to-agent path from lightweight chat. - [Realtime and canonicalization](https://useagent.org/docs/architecture/realtime): The two SSE streams, the canonicalization outbox, and why exactly one backend runs per database today. - [Shared packages](https://useagent.org/docs/architecture/shared-packages): The runtime-neutral contracts that let the frontend and backend agree without depending on each other's internals. ## Platform - [Platform](https://useagent.org/docs/platform): The capabilities that span more than one surface. Credentials and models, GitHub, usage accounting, and the tenant boundary. - [Provider connections and models](https://useagent.org/docs/platform/provider-connections-and-models): Two ways to authenticate an engine, the gateway providers behind them, and the model families served, with the credential never entering a sandbox. - [GitHub integration](https://useagent.org/docs/platform/github-integration): How repositories are selected, authenticated with a GitHub App, cloned server-side, and read by agents in a run. - [Fleet and usage](https://useagent.org/docs/platform/fleet-and-usage): Where the dashboard's usage numbers come from. Real model burn from run accounting, and a cached view of the sandbox fleet. - [Auth and organizations](https://useagent.org/docs/platform/auth-and-orgs): The tenant boundary. How sign-in works, how every resource is scoped to an organization, and how the dev-mode switches behave. ## Operations - [Operations](https://useagent.org/docs/operations): How UseAgent ships, how it proves a release, and the invariants production runs on. - [Deployment lanes](https://useagent.org/docs/operations/deployment): Two lanes with different jobs. A full certification and a source-convergence bootstrap, both ending in the same atomic frontend swap. - [Verification](https://useagent.org/docs/operations/verification): The test and parity arsenal, from fast isolated-database suites to real engine journeys run against a release candidate. - [Operational invariants](https://useagent.org/docs/operations/invariants): The properties production depends on. Break one and the guarantees above it stop holding. - [Backups and restore](https://useagent.org/docs/operations/backups): A daily encrypted snapshot of the database and durable state, with verify and restore lanes and an explicit off-host requirement. ## Channels - [Channels](https://useagent.org/docs/channels): Why a run behaves the same whether it starts in the web app, in Slack, or from a schedule. One durable command lane, one door. - [Slack](https://useagent.org/docs/channels/slack): Slack as a first-class, bidirectional channel. Mention to run, threaded replies, inbound attachments, repo binding from links, and results back in the thread. ## API reference - [API reference](https://useagent.org/docs/api): The HTTP surface the frontend and integrations call - the auth model, release-compatibility headers, streaming endpoints, and how the reference is organized. - [Runs, threads and realtime API](https://useagent.org/docs/api/runs): Create and cancel runs, stream a thread over SSE, answer questions and approvals, and reach the sandbox through the terminal, live, and desktop transports. - [Artifacts and uploads API](https://useagent.org/docs/api/artifacts): Durable run artifacts, workpiece read/edit/export, agent-proposed revisions, PDF page operations, and user uploads. - [Knowledge and wiki API](https://useagent.org/docs/api/knowledge): Knowledge records and hybrid search, wiki documents and revisions, learning-lane knowledge drafts, and repo-wiki generation. - [Memory API](https://useagent.org/docs/api/memory): The Memory Hub - pool search and browse, item correction and deletion, the capture outbox, and the recall ledger. - [Skills, tasks and projects API](https://useagent.org/docs/api/skills): Skills and playbooks CRUD, running a skill, GitHub skill import, learning-lane skill proposals, and the durable task board. - [Automations, secrets and keys API](https://useagent.org/docs/api/automations): Scheduled automations, write-only org secrets, and session-only API keys. - [Platform and integrations API](https://useagent.org/docs/api/platform): GitHub repos and pulls, user provider connections, tenant integrations, Slack ingress, the dashboard, and fleet usage.