UNPKG

openclaw

Version:

Multi-channel AI gateway with extensible messaging integrations

316 lines (262 loc) 19.2 kB
--- summary: "How OpenClaw separates model providers, models, channels, and agent runtimes" title: "Agent runtimes" read_when: - You are choosing between OpenClaw, Codex, ACP, or another native agent runtime - You are confused by provider/model/runtime labels in status or config - You are documenting support parity for a native harness --- An **agent runtime** owns one prepared model loop: it receives the prompt, drives model output, handles native tool calls, and returns the finished turn to OpenClaw. Runtimes are easy to confuse with providers because both show up near model configuration. They are different layers: | Layer | Examples | Meaning | | ------------- | -------------------------------------------- | ------------------------------------------------------------------- | | Provider | `anthropic`, `github-copilot`, `openai` | How OpenClaw authenticates, discovers models, and names model refs. | | Model | `claude-opus-4-6`, `gpt-5.6-sol` | The model selected for the agent turn. | | Agent runtime | `claude-cli`, `codex`, `copilot`, `openclaw` | The low-level loop or backend that executes the prepared turn. | | Channel | Discord, Slack, Telegram, WhatsApp | Where messages enter and leave OpenClaw. | A **harness** is the implementation that provides an agent runtime (code term). For example, the bundled Codex harness implements the `codex` runtime. Public config uses `agentRuntime.id` on provider or model entries; whole-agent runtime keys are legacy and ignored. `openclaw doctor --fix` removes old whole-agent runtime pins and rewrites legacy runtime model refs to canonical provider/model refs plus model-scoped runtime policy where needed. Two runtime families: - **Embedded harnesses** run inside OpenClaw's prepared agent loop: the built-in `openclaw` runtime, plus registered plugin harnesses such as `codex` and `copilot`. - **CLI backends** run a local CLI process while keeping the model ref canonical. For example, `anthropic/claude-opus-5` with a model-scoped `agentRuntime.id: "claude-cli"` means "select the Anthropic model, execute through Claude CLI." `claude-cli` is not an embedded harness id and must not be passed to AgentHarness selection. The `copilot` harness is a separate, opt-in external plugin harness for the GitHub Copilot CLI; see [GitHub Copilot agent runtime](/plugins/copilot) for the user-facing decision between PI, Codex, and GitHub Copilot agent runtime. ## Codex surfaces Several surfaces share the Codex name: | Surface | OpenClaw name/config | What it does | | ------------------------------------------------ | ------------------------------------ | -------------------------------------------------------------------------------------------------------------- | | Native Codex app-server runtime | `openai/*` model refs | Runs OpenAI embedded agent turns through Codex app-server. This is the usual ChatGPT/Codex subscription setup. | | Codex OAuth auth profiles | `openai` OAuth profiles | Stores ChatGPT/Codex subscription auth that the Codex app-server harness consumes. | | Codex ACP adapter | `runtime: "acp"`, `agentId: "codex"` | Runs Codex through the external ACP/acpx control plane. Use only when ACP/acpx is explicitly asked for. | | Native Codex chat-control command set | `/codex ...` | Binds, resumes, steers, stops, and inspects Codex app-server threads from chat. | | OpenAI Platform API route for non-agent surfaces | `openai/*` plus API-key auth | Direct OpenAI APIs such as images, embeddings, speech, and realtime. | These surfaces are intentionally independent. Enabling the `codex` plugin makes native app-server features available; `openclaw doctor --fix` owns legacy Codex route repair and stale session pin cleanup. Automatic Codex selection requires a compatible effective route: an exact official HTTPS Platform Responses or ChatGPT Responses endpoint without authored request overrides. The `openai/*` prefix alone does not select Codex. The common ChatGPT/Codex subscription setup uses Codex OAuth for auth, but keeps the model ref as `openai/*` and selects the `codex` runtime: ```json5 { agents: { defaults: { model: "openai/gpt-5.6-sol", }, }, } ``` That means OpenClaw selects an OpenAI model ref, then asks the Codex app-server runtime to run the embedded agent turn. It does not mean "use API billing," and it does not mean the channel, model provider catalog, or OpenClaw session store becomes Codex. When the bundled `codex` plugin is enabled, use the native `/codex` command surface (`/codex bind`, `/codex threads`, `/codex resume`, `/codex steer`, `/codex stop`) for natural-language Codex control instead of ACP. Use ACP for Codex only when the user explicitly asks for ACP/acpx or is testing the ACP adapter path. Claude Code, Gemini CLI, OpenCode, Cursor, and similar external harnesses still use ACP. Decision tree: 1. **Codex bind/control/thread/resume/steer/stop** -> native `/codex` command surface when the bundled `codex` plugin is enabled. 2. **Codex as the embedded runtime** or the normal subscription-backed Codex agent experience -> `openai/<model>`. 3. **OpenClaw explicitly chosen for an OpenAI model** -> keep the model ref as `openai/<model>` and set provider/model runtime policy to `agentRuntime.id: "openclaw"`. A selected `openai` OAuth profile is routed internally through OpenClaw's Codex-auth transport. 4. **Legacy Codex model refs in config** -> repair with `openclaw doctor --fix` to `openai/<model>`; doctor keeps the Codex auth route by adding provider/model-scoped `agentRuntime.id: "codex"` where the old model ref implied it. Legacy **`codex-cli/*`** model refs repair to the same `openai/<model>` Codex app-server route; OpenClaw no longer keeps a bundled Codex CLI backend. 5. **ACP, acpx, or Codex ACP adapter explicitly requested** -> `runtime: "acp"` and `agentId: "codex"`. 6. **Claude Code, Gemini CLI, OpenCode, Cursor, Droid, or another external harness** -> ACP/acpx, not the native sub-agent runtime. | You mean... | Use... | | --------------------------------------- | -------------------------------------------- | | Codex app-server chat/thread control | `/codex ...` from the bundled `codex` plugin | | Codex app-server embedded agent runtime | `openai/*` agent model refs | | OpenAI Codex OAuth | `openai` OAuth profiles | | Claude Code or other external harness | ACP/acpx | For the OpenAI-family prefix split, see [OpenAI](/providers/openai) and [Model providers](/concepts/model-providers). For the Codex runtime support contract, see [Codex harness runtime](/plugins/codex-harness-runtime#v1-support-contract). ## Runtime ownership Different runtimes own different amounts of the loop: | Surface | OpenClaw embedded | Codex app-server | | --------------------------- | ---------------------------------------------- | --------------------------------------------------------------------------- | | Model loop owner | OpenClaw, through the OpenClaw embedded runner | Codex app-server | | Canonical thread state | OpenClaw transcript | Codex thread, plus OpenClaw transcript mirror | | OpenClaw dynamic tools | Native OpenClaw tool loop | Bridged through the Codex adapter | | Native shell and file tools | OpenClaw path | Codex-native tools, bridged through native hooks where supported | | Context engine | Native OpenClaw context assembly | OpenClaw projects assembled context into the Codex turn | | Compaction | OpenClaw or selected context engine | Codex-native compaction, with OpenClaw notifications and mirror maintenance | | Channel delivery | OpenClaw | OpenClaw | Design rule: if OpenClaw owns the surface, it can provide normal plugin hook behavior. If the native runtime owns the surface, OpenClaw needs runtime events or native hooks. If the native runtime owns canonical thread state, OpenClaw mirrors and projects context rather than rewriting unsupported internals. A **locked concrete model chat** still uses normal model discovery, credential selection, and its configured request transport. The lock prevents model changes; it does not hand model or authentication ownership to a native runtime. Responses parameters and other authored request settings remain part of the concrete request. The selected runtime must support them or declare a lossless fallback before execution. A **bound native session** can instead retain its native model and, separately, its native connection's authentication. OpenClaw verifies that ownership against the exact pinned harness and its private binding, not a previous usage report. For Codex, native authentication belongs only to the separate supervision connection; preserving a native model on a managed connection still uses host auth preparation. Native-auth connections keep their own connection policy and do not receive a forwarded host profile. They reject explicit per-run provider stream parameters rather than silently dropping them. Use a concrete model chat when you need to apply those parameters. When a native model still uses host authentication, its actual provider/model pair controls credential and request preparation, not the outer default. An explicit auth profile remains locked. If resume changes that pair after credentials were prepared, the turn stops before inference and preserves the newly observed native state; it does not retry with stale credentials or replace the thread. Session rows and events use the native owner's known model pair. A pending native branch can show a configured placeholder until its first turn selects a model. For native-auth sessions, chat metadata removes the unrelated host-credential gate from that rendered row without claiming native credentials are ready. Global model availability and concrete-model chats keep their normal host-auth checks. The native selection also remains separate from a final response's billing model, including when a host finalizer supplies the last answer. ## Runtime selection OpenClaw resolves an embedded runtime after provider and model resolution, in this order: 1. **Model-scoped runtime policy** wins. This lives in a configured provider model entry, or in `agents.defaults.models["provider/model"].agentRuntime` / `agents.entries.*.models["provider/model"].agentRuntime`. A provider wildcard such as `agents.defaults.models["vllm/*"].agentRuntime` applies after exact model policy, so dynamically discovered provider models can share one runtime without overriding exact per-model exceptions. 2. **Provider-scoped runtime policy**: `models.providers.<provider>.agentRuntime`. 3. **`auto` mode**: registered plugin runtimes can claim supported provider/model pairs. 4. If nothing claims the turn in `auto` mode, OpenClaw falls back to `openclaw` as the compatibility runtime. Use an explicit runtime id when the run must be strict. Historical `agentHarnessId` records which runtime produced the transcript; it does not pin the next turn. An explicit trusted `pluginOwnerId` remains the session's control owner even after another harness reports usage. A model lock on that plugin-owned chat does not turn the observation into a native harness pin. Locked native transcripts retain their creating harness, and compatible explicit session runtime overrides take precedence over configured policy. ACP sessions retain their ACP backend. Legacy whole-agent runtime config and `OPENCLAW_AGENT_RUNTIME` are ignored; use `openclaw doctor --fix` to remove stale config and repair legacy model refs. Explicit provider/model plugin runtimes fail closed when the harness is missing or cannot support the route or authentication. There is one selection-time exception: a harness may declare that OpenClaw can reproduce the exact request. Codex uses this fallback for authored request overrides such as headers, request parameters, timeouts, or payload compatibility switches. It preserves those settings instead of silently dropping them. Once a harness starts executing, its failures are not replayed through another runtime. Affirmative `compat.supportsReasoningEffort: true` and a nonempty `compat.supportedReasoningEfforts` list containing only `minimal`, `low`, `medium`, `high`, `xhigh`, `max`, or `ultra` describe native reasoning capabilities; they do not opt an otherwise compatible route out of Codex. Disabling reasoning, custom effort labels, and other compatibility switches remain request behavior. Model-level runtime controls such as `fastMode` and `thinking` also preserve native selection when their values are valid. CLI backend aliases differ from embedded harness ids. Preferred Claude CLI form: ```json5 { agents: { defaults: { model: "anthropic/claude-opus-5", models: { "anthropic/claude-opus-5": { agentRuntime: { id: "claude-cli" }, }, }, }, }, } ``` Legacy refs such as `claude-cli/claude-opus-4-7` are accepted as compatibility input, but new config should keep the provider/model canonical and put the execution backend in provider/model runtime policy. Run `openclaw doctor --fix` to rewrite persisted legacy model selections, model-map keys, and explicit `modelPolicy.allow` entries to that canonical shape. Legacy `codex-cli/*` refs are different: doctor migrates them to `openai/*` so they run through the Codex app-server harness instead of preserving a Codex CLI backend. For OpenAI agent models, unset runtime and `auto` can select Codex when the provider-owned effective route declares it compatible. Custom endpoints, Completions adapters, and authored request overrides stay on OpenClaw rather than losing their transport settings. Explicit `agentRuntime.id: "openclaw"` also keeps the built-in runtime available; with a selected `openai` OAuth profile, it uses OpenClaw's Codex-auth transport while keeping the public model ref as `openai/*`. Stale historical producer fields do not pin the next turn and can be cleaned with `openclaw doctor --fix`. If `openclaw doctor` warns that the `codex` plugin is enabled while legacy Codex model refs remain in config, treat that as legacy route state and run `openclaw doctor --fix` to rewrite it to `openai/*` with the Codex runtime. ## GitHub Copilot agent runtime The external `@openclaw/copilot` plugin registers an opt-in `copilot` runtime backed by the GitHub Copilot CLI (`@github/copilot-sdk`). It claims the canonical subscription `github-copilot` provider and is **never** selected by `auto`. Opt in per-model or per-provider via `agentRuntime.id`: ```json5 { agents: { defaults: { model: "github-copilot/gpt-5.5", models: { "github-copilot/gpt-5.5": { agentRuntime: { id: "copilot" }, }, }, }, }, } ``` The plugin manifest declares the harness provider, runtime, CLI session key, and auth profile prefix without requiring `openclaw doctor` to load plugin code. For configuration, auth, transcript mirroring, compaction, the declarative doctor contract, and the broader PI vs Codex vs Copilot SDK decision, see [GitHub Copilot agent runtime](/plugins/copilot). ## Compatibility contract When a runtime is not OpenClaw, its docs should state which OpenClaw surfaces it supports: | Question | Why it matters | | -------------------------------------- | ------------------------------------------------------------------------------------------------- | | Who owns the model loop? | Determines where retries, tool continuation, and final answer decisions happen. | | Who owns canonical thread history? | Determines whether OpenClaw can edit history or only mirror it. | | Do OpenClaw dynamic tools work? | Messaging, sessions, cron, and OpenClaw-owned tools rely on this. | | Do dynamic tool hooks work? | Plugins expect `before_tool_call`, `after_tool_call`, and middleware around OpenClaw-owned tools. | | Do native tool hooks work? | Shell, patch, and runtime-owned tools need native hook support for policy and observation. | | Does the context engine lifecycle run? | Memory and context plugins depend on assemble, ingest, after-turn, and compaction lifecycle. | | What compaction data is exposed? | Some plugins only need notifications; others need kept/dropped metadata. | | What is intentionally unsupported? | Users should not assume OpenClaw equivalence where the native runtime owns more state. | The Codex runtime support contract is documented in [Codex harness runtime](/plugins/codex-harness-runtime#v1-support-contract). ## Status labels Status output can show both `Execution` and `Runtime` labels. Read them as diagnostics, not provider names: - A model ref such as `openai/gpt-5.6-sol` is the selected provider/model. - A runtime id such as `codex` is the loop executing the turn. - A channel label such as Telegram or Discord is where the conversation is happening. If a run shows an unexpected runtime, inspect the selected provider/model runtime policy first. Next-turn runtime metadata includes declared fallback when the registered harness can determine it from the configured route. It does not probe credentials or start a runtime; final route/auth preparation can still reject the turn. The completed result records the runtime that actually ran. ## Related - [Codex harness](/plugins/codex-harness) - [Codex harness runtime](/plugins/codex-harness-runtime) - [GitHub Copilot agent runtime](/plugins/copilot) - [OpenAI](/providers/openai) - [Agent harness plugins](/plugins/sdk-agent-harness) - [Agent loop](/concepts/agent-loop) - [Models](/concepts/models) - [Status](/cli/status)