What causes "Z.AI API Key must be provided to use agents." in Mintplex-Labs/anything-llm?

Create an API key in the Z.AI console (open.bigmodel.cn / z.ai) and set ZAI_API_KEY in .env.

Mintplex-Labs/anything-llm · error · Error

Z.AI API Key must be provided to use agents.

Error message

Z.AI API Key must be provided to use agents.

What it means

Thrown by AgentHandler.checkSetup() (server/utils/agents/index.js:222) when the workspace agent provider is 'zai' but ZAI_API_KEY is unset/empty. Z.AI (GLM family) exposes its API behind this key, which is the only credential the guard validates.

Source

Thrown at server/utils/agents/index.js:223

          throw new Error("DeepSeek API Key must be provided to use agents.");
        break;
      case "litellm":
        if (!process.env.LITE_LLM_BASE_PATH)
          throw new Error(
            "LiteLLM API base path and key must be provided to use agents."
          );
        break;
      case "apipie":
        if (!process.env.APIPIE_LLM_API_KEY)
          throw new Error("ApiPie API Key must be provided to use agents.");
        break;
      case "xai":
        if (!process.env.XAI_LLM_API_KEY)
          throw new Error("xAI API Key must be provided to use agents.");
        break;
      case "zai":
        if (!process.env.ZAI_API_KEY)
          throw new Error("Z.AI API Key must be provided to use agents.");
        break;
      case "novita":
        if (!process.env.NOVITA_LLM_API_KEY)
          throw new Error("Novita API Key must be provided to use agents.");
        break;
      case "nvidia-nim":
        if (!process.env.NVIDIA_NIM_LLM_BASE_PATH)
          throw new Error(
            "NVIDIA NIM base path must be provided to use agents."
          );
        break;
      case "ppio":
        if (!process.env.PPIO_API_KEY)
          throw new Error("PPIO API Key must be provided to use agents.");
        break;
      case "gemini":
        if (!process.env.GEMINI_API_KEY)
          throw new Error("Gemini API key must be provided to use agents.");

View on GitHub (pinned to 526360e320)

Solutions

Create an API key in the Z.AI console (open.bigmodel.cn / z.ai) and set ZAI_API_KEY in .env.
Restart AnythingLLM to reload process.env.
Verify the key with a curl to the Z.AI models endpoint using Authorization: Bearer $ZAI_API_KEY.
Confirm the workspace agent model is a valid Z.AI/GLM model id.

Example fix

// before (.env)
// ZAI_API_KEY=

// after (.env)
ZAI_API_KEY=...your-zai-key

Defensive patterns

Strategy: validation

Validate before calling

const key = process.env.ZAI_API_KEY;
if (!key) {
  throw new Error(
    'Cannot start Z.AI agent: set ZAI_API_KEY (from the Z.AI console) in .env.'
  );
}

Try / catch

try {
  await agentHandler.start();
} catch (e) {
  if (/Z\.AI API Key must be provided/.test(e.message)) {
    return respondConfigError('zai', ['ZAI_API_KEY']);
  }
  throw e;
}

Prevention

Create the key in the Z.AI console and write it to .env immediately.
Restart the Node process / Docker container after .env edits.
Add a CI/startup assertion that ZAI_API_KEY is set when provider is zai.
Keep .env out of version control.

When it happens

Trigger: An agent invocation runs with workspace.agentProvider === 'zai' while ZAI_API_KEY is absent from process.env. Path: #providerSetupAndCheck() -> checkSetup() at index.js:476.

Common situations: Z.AI selected in System > LLM Preference without a key; key from z.ai management console not pasted into .env; .env changed but AnythingLLM not restarted; key expired/revoked.

Related errors

AI-assisted analysis of Mintplex-Labs/anything-llm@526360e320 (2026-08-13). Data as JSON: /api/errors/3c03a4548e83e23b. Report an issue: GitHub.