Mintplex-Labs/anything-llm · warning · Error

Could not find indexes.

Error message

Could not find indexes.

What it means

Thrown by System.totalIndexes when the GET to {fullApiUrl()}/system/system-vectors (optionally ?slug=...) returns non-2xx. It reads the vector database index count for a workspace or instance. The .catch() returns 0, so a failure is indistinguishable from a genuinely empty vector store.

Source

Thrown at frontend/src/models/system.js:29

    supportEmail: "anythingllm_support_email",
    customAppName: "anythingllm_custom_app_name",
    canViewChatHistory: "anythingllm_can_view_chat_history",
    deploymentVersion: "anythingllm_deployment_version",
  },
  ping: async function () {
    return await fetch(`${API_BASE}/ping`)
      .then((res) => res.json())
      .then((res) => res?.online || false)
      .catch(() => false);
  },
  totalIndexes: async function (slug = null) {
    const url = new URL(`${fullApiUrl()}/system/system-vectors`);
    if (!!slug) url.searchParams.append("slug", encodeURIComponent(slug));
    return await fetch(url.toString(), {
      headers: baseHeaders(),
    })
      .then((res) => {
        if (!res.ok) throw new Error("Could not find indexes.");
        return res.json();
      })
      .then((res) => res.vectorCount)
      .catch(() => 0);
  },

  /**
   * Checks if the onboarding is complete.
   * @returns {Promise<boolean>}
   */
  isOnboardingComplete: async function () {
    return await fetch(`${API_BASE}/onboarding`)
      .then((res) => {
        if (!res.ok) throw new Error("Could not find onboarding information.");
        return res.json();
      })
      .then((res) => res.onboardingComplete)
      .catch(() => false);

View on GitHub (pinned to 526360e320)

Solutions

1. Check server logs for the /system/system-vectors handler, which surfaces the underlying vectorDB connection error.
2. Verify VectorDatabase provider env vars are set and the embedding engine is selected in system settings.
3. If passing a slug, confirm it matches a workspace returned by the workspaces listing endpoint.
4. Distinguish a real 0 from an error by temporarily removing the .catch(() => 0) during debugging.

Example fix

// before
const count = await System.totalIndexes(slug); // 0 on error OR empty

// after (debug: let it throw to see the real cause)
const res = await fetch(`${fullApiUrl()}/system/system-vectors` + (slug ? `?slug=${encodeURIComponent(slug)}` : ""), { headers: baseHeaders() });
if (!res.ok) throw new Error(`vectors ${res.status}`);
const count = (await res.json()).vectorCount;

// Validate slug shape if provided.
function validSlug(slug) {
  return slug == null || (typeof slug === "string" && /^[\w-:]+$/.test(slug));
}

/** @param {any} n @returns {n is number} */
function isNonNegativeInt(n) { return Number.isInteger(n) && n >= 0; }

const count = await System.totalIndexes(slug);
const safe = isNonNegativeInt(count) ? count : 0; // 0 may be error OR empty

When it happens

Trigger: Calling totalIndexes() or totalIndexes("my-workspace") when the vector database is unreachable/misconfigured (500), when the slug does not match any workspace, or when the auth token is rejected (401/403).

Common situations: Vector DB provider (Pinecone/Chroma/Qdrant/Weaviate) credentials are unset or wrong in env; the embedding engine is configured but the vector DB is not; the workspace slug was renamed and the stale slug is passed.

Related errors

• Failed to sync Confluence page content. ${reason}
• Failed to sync GitHub file content. ${reason}
• Failed to sync GitLab file content. ${reason}
• Failed to sync Gitea file content. ${reason}
• Failed to fetch document content

AI-assisted analysis of Mintplex-Labs/anything-llm@526360e320 (2026-08-13). Data as JSON: /api/errors/c9623e9e192b5134. Report an issue: GitHub.