tinyhumansai/openhuman · error · Error

Not running in Tauri

Error message

Not running in Tauri

What it means

openhumanAgentChat sends a one-shot chat turn to the core via openhuman.agent_chat ({ message, model_override, temperature }) and is guarded by the shared isTauri() check that throws 'Not running in Tauri'. The chat path only exists through relay_http_rpc (a Tauri command invoked over IPC), so outside the desktop shell — or before window.__TAURI_INTERNALS__.invoke is injected at CEF startup (#1472) — the wrapper fails fast with this error instead of attempting invoke.

Source

Thrown at app/src/utils/tauriCommands/localAi.ts:230

    chat_eligibility?: ModelContextEligibility | null;
    embedding_model: string;
    embedding_found: boolean;
    embedding_eligibility?: ModelContextEligibility | null;
    vision_model: string;
    vision_found: boolean;
  };
  issues: string[];
  repair_actions: RepairAction[];
  ok: boolean;
}

export async function openhumanAgentChat(
  message: string,
  modelOverride?: string,
  temperature?: number
): Promise<CommandResponse<string>> {
  if (!isTauri()) {
    throw new Error('Not running in Tauri');
  }
  return await callCoreRpc<CommandResponse<string>>({
    method: 'openhuman.agent_chat',
    params: { message, model_override: modelOverride, temperature },
  });
}

export async function openhumanLocalAiStatus(): Promise<CommandResponse<LocalAiStatus>> {
  try {
    return await callCoreRpc<CommandResponse<LocalAiStatus>>({
      method: 'openhuman.inference_status',
    });
  } catch (err) {
    const message = tauriErrorMessage(err);
    if (message.includes('unknown method: openhuman.inference_status')) {
      throw new Error(
        'Local model runtime is unavailable in this core build. Restart app after updating to the latest build.'
      );

View on GitHub (pinned to 7491200858)

Solutions

  1. Run chats inside the Tauri host: pnpm dev:app.
  2. Mock the Tauri core module (isTauri → true, invoke → canned CommandResponse<string>) in tests.
  3. Route non-Tauri contexts through the regular web-chat service instead of this desktop-only wrapper.
  4. Gate agent-chat debug actions on isTauri() and core readiness.

Example fix

// before
const reply = (await openhumanAgentChat('ping')).result;

// after
if (!isTauri()) throw new Error('agent_chat requires the desktop runtime');
const reply = (await openhumanAgentChat('ping')).result;
Defensive patterns

Strategy: validation

Validate before calling

if (!isTauri()) {
  throw new Error('agent_chat requires the desktop runtime');
}
const reply = (await openhumanAgentChat(message)).result;

Type guard

import { isTauri } from './common';
const agentChatAvailable = (): boolean => isTauri();

Try / catch

try { const reply = (await openhumanAgentChat(msg)).result; }
catch (e) {
  if (e instanceof Error && e.message === 'Not running in Tauri') { useWebChatFallback(); return; }
  throw e;
}

Prevention

When it happens

Trigger: Calling openhumanAgentChat('hello') from the Vite dev server in a normal browser; a chat unit test that doesn't mock the Tauri API; a debug console/panel issuing a test message during the startup bridge gap.

Common situations: Using pnpm dev instead of pnpm dev:app while iterating on chat; integration tests of chat plumbing missing '@tauri-apps/api/core' mocks; early-mount debug tools firing before IPC readiness.

Related errors

AI-assisted analysis of tinyhumansai/openhuman@7491200858 (2026-08-17). Data as JSON: /api/errors/949de14d37914a7d. Report an issue: GitHub.