The bundled openai plugin registers speech synthesis for the messages.tts surface.

Available models: gpt-4o-mini-tts, tts-1, tts-1-hd. Available voices: alloy, ash, ballad, cedar, coral, echo, fable, juniper, marin, onyx, nova, sage, shimmer, verse.

{
  messages: {
    tts: {
      providers: {
        openai: { model: "gpt-4o-mini-tts", voice: "coral" },
      },
    },
  },
}

The bundled openai plugin registers realtime transcription for the Voice Call plugin.

The bundled openai plugin registers realtime voice for the Voice Call plugin.

OpenClaw uses WebSocket-first with SSE fallback ("auto") for both openai/* and openai-codex/*.

In "auto" mode, OpenClaw:

• Retries one early WebSocket failure before falling back to SSE
• After a failure, marks WebSocket as degraded for ~60 seconds and uses SSE during cool-down
• Attaches stable session and turn identity headers for retries and reconnects
• Normalizes usage counters (input_tokens / prompt_tokens) across transport variants

{
  agents: {
    defaults: {
      models: {
        "openai-codex/gpt-5.4": {
          params: { transport: "auto" },
        },
      },
    },
  },
}

Related OpenAI docs:

• Realtime API with WebSocket
• Streaming API responses (SSE)

OpenClaw enables WebSocket warm-up by default for openai/* to reduce first-turn latency.

// Disable warm-up
{
  agents: {
    defaults: {
      models: {
        "openai/gpt-5.4": {
          params: { openaiWsWarmup: false },
        },
      },
    },
  },
}

OpenClaw exposes a shared fast-mode toggle for both openai/* and openai-codex/*:

• Chat/UI: /fast status|on|off
• Config: agents.defaults.models["<provider>/<model>"].params.fastMode

When enabled, OpenClaw maps fast mode to OpenAI priority processing (service_tier = "priority"). Existing service_tier values are preserved, and fast mode does not rewrite reasoning or text.verbosity.

{
  agents: {
    defaults: {
      models: {
        "openai/gpt-5.4": { params: { fastMode: true } },
        "openai-codex/gpt-5.4": { params: { fastMode: true } },
      },
    },
  },
}

OpenAI's API exposes priority processing via service_tier. Set it per model in OpenClaw:

{
  agents: {
    defaults: {
      models: {
        "openai/gpt-5.4": { params: { serviceTier: "priority" } },
        "openai-codex/gpt-5.4": { params: { serviceTier: "priority" } },
      },
    },
  },
}

Supported values: auto, default, flex, priority.

For direct OpenAI Responses models (openai/* on api.openai.com), OpenClaw auto-enables server-side compaction:

• Forces store: true (unless model compat sets supportsStore: false)
• Injects context_management: [{ type: "compaction", compact_threshold: ... }]
• Default compact_threshold: 70% of contextWindow (or 80000 when unavailable)

{
  agents: {
    defaults: {
      models: {
        "azure-openai-responses/gpt-5.4": {
          params: { responsesServerCompaction: true },
        },
      },
    },
  },
}

{
  agents: {
    defaults: {
      models: {
        "openai/gpt-5.4": {
          params: {
            responsesServerCompaction: true,
            responsesCompactThreshold: 120000,
          },
        },
      },
    },
  },
}

{
  agents: {
    defaults: {
      models: {
        "openai/gpt-5.4": {
          params: { responsesServerCompaction: false },
        },
      },
    },
  },
}

For GPT-5-family runs on openai/* and openai-codex/*, OpenClaw can use a stricter embedded execution contract:

{
  agents: {
    defaults: {
      embeddedPi: { executionContract: "strict-agentic" },
    },
  },
}

With strict-agentic, OpenClaw:

• No longer treats a plan-only turn as successful progress when a tool action is available
• Retries the turn with an act-now steer
• Auto-enables update_plan for substantial work
• Surfaces an explicit blocked state if the model keeps planning without acting

OpenClaw treats direct OpenAI, Codex, and Azure OpenAI endpoints differently from generic OpenAI-compatible /v1 proxies:

Native routes (openai/*, openai-codex/*, Azure OpenAI):

• Keep reasoning: { effort: "none" } intact when reasoning is explicitly disabled
• Default tool schemas to strict mode
• Attach hidden attribution headers on verified native hosts only
• Keep OpenAI-only request shaping (service_tier, store, reasoning-compat, prompt-cache hints)

Proxy/compatible routes:

• Use looser compat behavior
• Do not force strict tool schemas or native-only headers

Azure OpenAI uses native transport and compat behavior but does not receive the hidden attribution headers.