Mintplex-Labs/anything-llm · error · Error

No valid provider found for the agent.

Error message

No valid provider found for the agent.

What it means

Thrown by EphemeralAgentHandler.#fetchModel() (server/utils/agents/ephemeral.js:187) when this.provider is null AND #getFallbackProvider() also returns null. The fallback path tries the system LLM_PROVIDER and calls providerDefault() to find a model; it returns null when LLM_PROVIDER is unset or set to a provider whose default model cannot be resolved. Net effect: neither the workspace nor the system provides a usable agent provider+model pair.

Source

Thrown at server/utils/agents/ephemeral.js:192

      };
    }

    return null;
  }

  /**
   * Finds or assumes the model preference value to use for API calls.
   * If multi-model loading is supported, we use their agent model selection of the workspace
   * If not supported, we attempt to fallback to the system provider value for the LLM preference
   * and if that fails - we assume a reasonable base model to exist.
   * @returns {string|null} the model preference value to use in API calls
   */
  #fetchModel() {
    // Provider was not explicitly set for workspace, so we are going to run our fallback logic
    // that will set a provider and model for us to use.
    if (!this.provider) {
      const fallback = this.#getFallbackProvider();
      if (!fallback) throw new Error("No valid provider found for the agent.");
      this.provider = fallback.provider; // re-set the provider to the fallback provider so it is not null.
      return fallback.model; // set its defined model based on fallback logic.
    }

    // The provider was explicitly set, so check if the workspace has an agent model set.
    if (this.#workspace?.agentModel) return this.#workspace.agentModel;

    // Otherwise, we have no model to use - so guess a default model to use via the provider
    // and it's system ENV params and if that fails - we return either a base model or null.
    return this.providerDefault();
  }

  async #providerSetupAndCheck() {
    this.provider = this.#workspace?.agentProvider ?? null;
    this.model = this.#fetchModel();

    // If provider resolved to model router, resolve the actual provider/model
    if (this.provider === "anythingllm-router") {

View on GitHub (pinned to 526360e320)

Solutions

1. Configure a system LLM provider in System Settings -> LLM Preference (sets LLM_PROVIDER) plus its model, then retry.
2. Or set the workspace Agent Provider and Agent Model explicitly.
3. For providers that need a model-pref env (OPEN_MODEL_PREF, MOONSHOT_AI_MODEL_PREF, etc.), set that env too so providerDefault() returns a value.

Example fix

// before: LLM_PROVIDER unset, workspace has no agentProvider
// after (.env)
LLM_PROVIDER=openai
OPEN_AI_API_KEY=sk-...
OPEN_MODEL_PREF=gpt-4o-mini

function canResolveEphemeralProvider(workspace) {
  if (workspace?.agentProvider) return true;
  const sys = process.env.LLM_PROVIDER;
  return Boolean(sys); // providerDefault() must also yield a model
}
if (!canResolveEphemeralProvider(workspace)) return showSetupError('Configure a system or workspace agent provider.');

try { await ephemeral.init(); } catch (e) { if (/No valid provider found for the agent/.test(e.message)) return promptProviderSetup(); throw e; }

When it happens

Trigger: Ephemeral agent init runs with #workspace.agentProvider unset/null, and process.env.LLM_PROVIDER is unset OR set to a provider whose providerDefault() yields no model.

Common situations: Fresh install where the system LLM preference was never set; workspace created before any provider configured; LLM_PROVIDER points at a provider that needs a model-pref env (e.g. OPEN_MODEL_PREF) which is also unset; multi-model loading disabled and no system default.

Related errors

• No valid provider found for the agent.
• Foundry base path must be provided to use agents.
• GiteeAI API Key must be provided to use agents.
• Cohere API key must be provided to use agents.
• Docker Model Runner base path must be provided to use agents

AI-assisted analysis of Mintplex-Labs/anything-llm@526360e320 (2026-08-13). Data as JSON: /api/errors/c73b81d51e49ce35. Report an issue: GitHub.