abhigyanpatwari/GitNexus · error · Error

No suitable device found

Error message

No suitable device found

What it means

Terminal failure of initEmbedder's device-probing loop: every candidate device failed for reasons that did not short-circuit earlier (download failures and the cpu-specific message throw their own errors), so no inference device could run the embedding model. The embedder instance and init promise are reset, so a later call retries the whole probe.

Source

Thrown at gitnexus/src/mcp/core/embedder.ts:160

          // Network errors and circuit-open errors are not device-specific —
          // they will fail the same way on every device. Rethrow immediately
          // with actionable HF_ENDPOINT guidance rather than silently falling
          // back to the next device.
          const errMsg = deviceError instanceof Error ? deviceError.message : String(deviceError);
          if (isHfDownloadFailure(errMsg)) {
            const endpointHint = process.env.HF_ENDPOINT
              ? `The configured endpoint (${process.env.HF_ENDPOINT}) may be unreachable.`
              : `huggingface.co may be unreachable from your network.\n` +
                `  Set HF_ENDPOINT to a mirror and retry:\n` +
                `    HF_ENDPOINT=https://hf-mirror.com npx gitnexus analyze --embeddings\n` +
                `    (Windows: set HF_ENDPOINT=https://hf-mirror.com && npx gitnexus analyze --embeddings)`;
            throw new Error(`Failed to download embedding model: ${errMsg}\n  ${endpointHint}`);
          }
          if (device === 'cpu') throw new Error('Failed to load embedding model');
        }
      }

      throw new Error('No suitable device found');
    } catch (error) {
      isInitializing = false;
      initPromise = null;
      embedderInstance = null;
      throw error;
    } finally {
      isInitializing = false;
    }
  })();

  return initPromise;
};

/**
 * Check if embedder is ready
 */
export const isEmbedderReady = (): boolean => isHttpMode() || embedderInstance !== null;

View on GitHub (pinned to aac7515d2a)

Solutions

1. Check for resource exhaustion first: free memory, watch dmesg/task manager for OOM kills, and retry on a less constrained host or container.
2. Pin a device explicitly with GITNEXUS_EMBEDDING_DEVICE=cpu to skip broken accelerated paths.
3. Reinstall the embedding stack via `gitnexus embeddings install` to repair a broken onnxruntime/transformers pairing.
4. Switch to HTTP embeddings with GITNEXUS_EMBEDDING_URL + GITNEXUS_EMBEDDING_MODEL so no local device is needed.
5. If all routes fail, capture the per-device errors from the logs and file a GitNexus issue.

Example fix

# before: every device candidate fails
$ gitnexus analyze --embeddings
# → No suitable device found

# after: pin cpu, repair stack, or move to HTTP embeddings
$ GITNEXUS_EMBEDDING_DEVICE=cpu gitnexus embeddings install
$ GITNEXUS_EMBEDDING_DEVICE=cpu gitnexus analyze --embeddings
# or: export GITNEXUS_EMBEDDING_URL=... GITNEXUS_EMBEDDING_MODEL=... and retry

// Pre-flight: memory headroom and pinned device before embedding runs
import os from 'node:os';

function embeddingPreconditionsOk(): boolean {
  const freeMb = os.freemem() / 1024 / 1024;
  return freeMb > 1024; // model load needs real headroom, not the bare minimum
}
process.env.GITNEXUS_EMBEDDING_DEVICE ??= 'cpu'; // skip broken accelerated paths by default

try {
  await initEmbedder();
} catch (err) {
  if (err instanceof Error && err.message === 'No suitable device found') {
    // all devices failed: degrade gracefully, index without semantic search
    logger.warn('embeddings unavailable — continuing without them');
    return runAnalyze({ embeddings: false });
  }
  throw err;
}

When it happens

Trigger: GPU device candidates fail (driver/runtime mismatch, no WebGL/WASM support in the Node build) AND the cpu attempt fails for a non-network reason such as memory exhaustion or a broken native runtime — the loop exhausts all devices without a classifiable error.

Common situations: Containers with strict memory limits where model load OOMs on every device; minimal Node runtimes missing WASM support; hosts with mismatched GPU drivers where both accelerated and fallback paths are broken; rarely, deeply corrupted installs.

Related errors

• embedding device must be one of auto, dml, cuda, cpu, wasm;
• Local semantic embeddings are unavailable on macOS Intel (da
• Local semantic embeddings are unavailable: the optional embe
• No suitable device found for embedding model
• Local semantic embeddings are unavailable on macOS Intel (da

AI-assisted analysis of abhigyanpatwari/GitNexus@aac7515d2a (2026-08-20). Data as JSON: /api/errors/5acabbf301d42ee1. Report an issue: GitHub.