> ## Documentation Index > Fetch the complete documentation index at: https://docs.obversa.ai/llms.txt > Use this file to discover all available pages before exploring further. # API Reference > Every export a user calls, the package it comes from, and the page that teaches it. Find the export you need, the package it comes from, and the page that shows it running. Every package needs Node.js 22.12 or later; the tool rows below say what else each one needs. ## Exports | Export | Package | What it's for | Page | | ------------------------------------------------------------------------------------------------------------------------------------------------ | ----------------------------------- | --------------------------------------- | ------------------------------------------------------------------ | | `workflow`, `stage`, `person`, `briefFromFile` | `@obversa/runtime` | Declare a team | [Runtime](/packages/runtime#declare-a-team) | | `formatEvent` | `@obversa/runtime` | Print one run event | [Runtime](/packages/runtime#declare-a-team) | | `pipeline`, `dag`, `sequence`, `parallel`, `loop` | `@obversa/runtime` | The shapes a workflow takes | [Runtime](/packages/runtime#build-a-graph) | | `fnJob`, `agentJob`, `commandJob`, `gateJob`, `prove` | `@obversa/runtime` | The steps | [Runtime](/packages/runtime#build-a-graph) | | `isolated`, `writeScope` | `@obversa/runtime` | A step in its own worktree | [Workspace](/concepts/workspace) | | `passed`, `failed`, `all`, `any`, `not`, `predicate`, `quorum` | `@obversa/runtime` | Conditions on a step | [A command decides the path](/patterns/command-kickback) | | `agentCheck`, `confidenceCondition`, `minConfidence` | `@obversa/runtime` | Scored conditions | [Evals](/reviewing/evals#score-a-condition) | | `reviewPanel` | `@obversa/runtime` | Several reviewers, a threshold | [A review panel with a threshold](/patterns/review-panel) | | `tournament` | `@obversa/runtime` | Candidates, a judge, one winner | [Three candidates, one winner](/patterns/tournament) | | `kickback`, `revisionRequest` | `@obversa/runtime` | Fail a step with findings | [Feedback loops](/concepts/feedback-loops) | | `approval` | `@obversa/runtime` | A person's decision as a step | [A person decides](/patterns/approval) | | `createCallbackGate`, `createCallbackClient`, `createStoredCallbackClient`, `replayCallbackClient`, `directRouter` | `@obversa/runtime` | A question with a typed answer | [Callback gates](/reviewing/callback-gates) | | `createApprovalCallbackGate`, `resolveApproval`, `writeProofArtifact`, `createAcceptedResultRecord`, `resolveAcceptedResult`, `createProofCache` | `@obversa/runtime` | Bind a decision to exact bytes | [Proof-bound acceptance and approval](/reviewing/proof-acceptance) | | `run` | `@obversa/runtime` | Run a job and record it | [Running](/concepts/running) | | `team`, `teamGraphType`, `projectTeamRooms` | `@obversa/runtime` | Named members who talk | [Team conversation](/patterns/team-conversation) | | `compileGraph`, `resolveGraphPlan`, `dagGraphType`, `convergence` | `@obversa/runtime` | Graph forms and plans | [Outside graph types](/graphs/contract) | | `persistRunDefinition`, `loadRunDefinition`, `createGraphExecutor` | `@obversa/runtime` | Store and execute a graph | [Graph executor](/graphs/executor) | | `readRunPreflight`, `interruptRunPreflight`, `preflight` | `@obversa/runtime` | Engine checks before the first step | [Graph executor](/graphs/executor) | | `createLocalRunStorage` | `@obversa/runtime/storage/local` | The on-disk event and artifact store | [Events and artifacts](/recording/events-and-artifacts) | | `createGitWorktreeProvider` | `@obversa/runtime` | Capture, verify, lease, fork | [Workspace contract](/workspace/contract) | | `ground`, `curate`, `consolidate` | `@obversa/runtime/memory` | Memory through your own functions | [Memory mechanics](/memory/mechanics) | | `MockEngine`, `defineGraphDefinition`, `runGraphTypeConformance` | `@obversa/runtime/testing` | Run offline, prove a form | [Runtime](/packages/runtime#api) | | `outcomeFromAgentText`, `assertDistinctSeats` | `@obversa/runtime/workflow-support` | Read a reviewer's reply, check families | [Runtime](/packages/runtime#api) | | `startSupervisedRun`, `resumeSupervisedRun`, `readSupervisedRunStatus` | `@obversa/runner` | A run under a watchdog | [Supervised local runs](/driving/runner) | | `validateAgentResult`, `finalResultPart`, `modelIdentity` | `@obversa/api` | The engine contract | [Safe node attempts](/recording/node-attempts) | | `runEngineConformance`, `runMemoryConformance` | `@obversa/api/testing` | Check an adapter | [API](/packages/api#api) | | `runChild` | `@obversa/core` | One child process to a deadline | [Core](/packages/core) | | `resolveCommandExecutable`, `runOwnedCommand` | `@obversa/core/command` | Find and run a tool | [Core](/packages/core#run-an-owned-command) | | `writerReviewerPair`, `thresholdPanel`, `featureDelivery` | `@obversa/builtin-workflows` | Ready-made teams | [Built-in Workflows](/packages/builtin-workflows) | | `runSurface`, `startSurface`, `openSurfaceUrl`, `parseFramedResult` | `@obversa/surface` | One page for one decision | [Surface](/packages/surface) | | `reviewDiff`, `parseUnifiedDiff`, `appendToThread` | `@obversa/surface-diff` | Review a diff in a browser | [Surface Diff](/packages/surface-diff) | | `claude`, `ClaudeCliEngine` | `@obversa/engine-claude-cli` | Claude Code as a seat | [Claude CLI Engine](/packages/engine-claude-cli) | | `codex`, `CodexEngine` | `@obversa/engine-codex-cli` | Codex as a seat | [Codex CLI Engine](/packages/engine-codex-cli) | | `GrokCliEngine` | `@obversa/engine-grok-cli` | Grok CLI as an engine | [Grok CLI Engine](/packages/engine-grok-cli) | | `opencode`, `OpenCodeCliEngine` | `@obversa/engine-opencode-cli` | OpenCode as a seat | [OpenCode CLI Engine](/packages/engine-opencode-cli) | | `AgentSdkEngine` | `@obversa/engine-claude-agent-sdk` | The Claude Agent SDK as an engine | [Claude Agent SDK Engine](/packages/engine-claude-agent-sdk) | | `AnthropicApiEngine` | `@obversa/engine-anthropic-api` | The Anthropic API as an engine | [Anthropic API Engine](/packages/engine-anthropic-api) | | `JevApiEngine` | `@obversa/engine-jev-api` | Typed questions about a run | [Jev API Engine](/packages/engine-jev-api) | | `openGitMemory`, `openReasoningRecord` | `@obversa/memory-git` | Memory in Git, reasoning in commits | [Git Memory](/packages/memory-git) | | `createSimpleMemory` | `@obversa/memory-simple` | Memory in one process | [Simple Memory](/packages/memory-simple) | | `openMarkdownCorpus` | `@obversa/memory-markdown` | Search your Markdown files | [Markdown Memory](/packages/memory-markdown) | | `webhookNotifier` | `@obversa/notify-webhook` | A message per run event | [Webhook Notifications](/packages/notify-webhook) | Import from the package named in the row: ```ts examples/teams/writer-reviewer-pair.ts (excerpt) theme={null} import { claude } from '@obversa/engine-claude-cli'; import { codex } from '@obversa/engine-codex-cli'; import { run } from '@obversa/runtime'; import { briefFromFile, stage, workflow, type TeamSeat } from '@obversa/runtime'; ``` ## Tools Each row says what the tool is, what you need installed, and where its home is. * **Claude Code.** `@obversa/engine-claude-cli` runs each request in a fresh Claude Code process. You need Claude Code installed and signed in. Home: [claude.com/code](https://claude.com/code). * **Codex CLI.** `@obversa/engine-codex-cli` runs each request in a fresh Codex CLI process. You need the Codex CLI installed and signed in. Home: [github.com/openai/codex](https://github.com/openai/codex). * **OpenCode CLI.** `@obversa/engine-opencode-cli` runs each request in a fresh OpenCode process with a clean environment, so your own sign-in isn't used. You need OpenCode 1.18.23 installed, and you pass your provider's credentials through the engine's `auth` or `environment` option. Home: [opencode.ai](https://opencode.ai). * **Grok CLI.** `@obversa/engine-grok-cli` runs each request in a fresh Grok CLI process with a clean environment, so your own sign-in isn't used. You need Grok CLI 1.0.5 installed, and you pass your credentials through the engine's `environment` or `authFile` option. Home: [docs.x.ai/build/cli/headless-scripting](https://docs.x.ai/build/cli/headless-scripting). * **Anthropic API.** `@obversa/engine-anthropic-api` sends each request to the Anthropic API. You need an Anthropic API key. Home: [console.anthropic.com](https://console.anthropic.com). * **Claude Agent SDK.** `@obversa/engine-claude-agent-sdk` runs each request through the Claude Agent SDK. You need Claude Code signed in on the same machine, because the SDK uses its sign-in. Home: [github.com/anthropics/claude-agent-sdk](https://github.com/anthropics/claude-agent-sdk). * **Jev API.** `@obversa/engine-jev-api` answers structured questions about a run's recorded state. It never reads or edits files. You need a Jev endpoint and an API key. See [Jev API Engine](/packages/engine-jev-api). * **In-process memory.** `@obversa/memory-simple` keeps memory inside one process. You need nothing else. Home: [Simple Memory](/packages/memory-simple). * **Git-backed memory.** `@obversa/memory-git` stores memory in private Git references, so nothing lands in the working tree, and can write a stage's reasoning into the commit that carries its change. You need Git installed. Home: [Git Memory](/packages/memory-git). * **Markdown search.** `@obversa/memory-markdown` selects ranked passages from a local folder of Markdown files and gives `ground` the selected files through a read-only memory view. You need nothing else. Home: [Markdown Memory](/packages/memory-markdown). * **Webhook notifications.** `@obversa/notify-webhook` posts one message per interesting run event to a URL you supply; the body carries a `text` field, so a Slack, Discord or Teams incoming webhook shows it with no code of its own. You need a URL to post to. Home: [Webhook Notifications](/packages/notify-webhook). * **The review command.** `@obversa/surface-diff` ships `obversa-review`, which opens a diff for inline review. You need Git. See [Review a diff in a host pane](/hosts/review). * **The cmux host.** The cmux glue opens surfaces as browser splits beside the terminal and one editor window with the two worktrees. You need cmux installed. Home: [Obversa in cmux](/hosts/cmux). * **Plannotator.** Plannotator opens its review page in a split through the `PLANNOTATOR_BROWSER` variable the workspace launcher sets. You need Plannotator installed. Home: [github.com/backnotprop/plannotator](https://github.com/backnotprop/plannotator). ## Packages | Package | Purpose | Version | | ---------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------- | ------- | | `@obversa/obversa` | Installs every package below except Jev | `0.1.0` | | `@obversa/runtime` | Builds and runs workflows as graphs, team conversations or a review loop | `0.1.0` | | `@obversa/runner` | Supervises stored graph runs in bounded workers | `0.1.0` | | `@obversa/api` | Defines every contract: engine, memory, graph, events, artifacts, workspace, callbacks, proofs and run definitions, with their validators | `0.1.0` | | `@obversa/core` | Runs one child process with a deadline and a readable result | `0.1.0` | | `@obversa/memory-simple` | Stores memory in one process | `0.1.0` | | `@obversa/memory-git` | Stores memory in private Git references, and writes a stage's reasoning into its commit | `0.1.0` | | `@obversa/memory-markdown` | Selects ranked passages from local Markdown files | `0.1.0` | | `@obversa/notify-webhook` | Posts one message per interesting run event to a URL you supply | `0.1.0` | | `@obversa/builtin-workflows` | Three ready-made workflows | `0.1.0` | | `@obversa/surface` | Runs one local surface session | `0.1.0` | | `@obversa/surface-diff` | Opens a git diff for inline review | `0.1.0` | | `@obversa/engine-claude-agent-sdk` | Runs one engine attempt through the Agent SDK | `0.1.0` | | `@obversa/engine-anthropic-api` | Runs one engine attempt through the Anthropic API | `0.1.0` | | `@obversa/engine-claude-cli` | Runs one engine attempt through the Claude CLI | `0.1.0` | | `@obversa/engine-codex-cli` | Runs one engine attempt through the Codex CLI | `0.1.0` | | `@obversa/engine-grok-cli` | Runs one engine attempt through the Grok CLI | `0.1.0` | | `@obversa/engine-opencode-cli` | Runs one engine attempt through the OpenCode CLI | `0.1.0` | | `@obversa/engine-jev-api` | Answers decision questions over recorded run state through the Jev API | `0.1.0` | The api, core, runtime and runner packages share one version. The runtime never imports an engine or a memory adapter: you pass the ones you want when you run. ## Pages **Packages.** [Obversa](/packages/obversa), [Runtime](/packages/runtime), [Runner](/packages/runner), [API](/packages/api), [Core](/packages/core), [Built-in Workflows](/packages/builtin-workflows), [Surface](/packages/surface), [Surface Diff](/packages/surface-diff). **Engines.** [Claude CLI](/packages/engine-claude-cli), [Codex CLI](/packages/engine-codex-cli), [Grok CLI](/packages/engine-grok-cli), [OpenCode CLI](/packages/engine-opencode-cli), [Claude Agent SDK](/packages/engine-claude-agent-sdk), [Anthropic API](/packages/engine-anthropic-api), [Jev API](/packages/engine-jev-api). Coming soon: [eve](/packages/engine-eve), [Mastra](/packages/engine-mastra), [LangGraph](/packages/engine-langgraph). **Memory and notifications.** [Git Memory](/packages/memory-git), [Simple Memory](/packages/memory-simple), [Markdown Memory](/packages/memory-markdown), [Webhook Notifications](/packages/notify-webhook). ## Next steps * [Installation](/get-started/installation): set a project up and run the first file. * [Runtime](/packages/runtime): the package most files import from. * [Examples](/workflows): complete teams that use these exports.