linshenkx/prompt-optimizer · warning · APIError

API returned empty model list

Error message

API returned empty model list

What it means

OpenAIAdapter.getModelsAsync successfully fetched a models array (data: [...]) but after filtering/mapping it produced zero entries, so it throws rather than return an empty list — the endpoint is reachable and the shape is right, but there is nothing usable.

Source

Thrown at packages/core/src/services/llm/adapters/openai-adapter.ts:159

    // 验证baseURL以/v1结尾
    const baseURL = config.connectionConfig.baseURL || this.getProvider().defaultBaseURL

    const openai = this.createOpenAIInstance(config, false)

    try {
      const response = await openai.models.list()

      // 检查返回格式
      if (response && response.data && Array.isArray(response.data)) {
        const models = response.data
          .map((model) => {
            // 使用buildDefaultModel为每个模型ID创建TextModel对象
            return this.buildDefaultModel(model.id)
          })
          .sort((a, b) => a.id.localeCompare(b.id))

        if (models.length === 0) {
          throw new APIError('API returned empty model list')
        }

        return models
      }

      throw new APIError('Unexpected API response format')
    } catch (error: any) {
      console.error('[OpenAIAdapter] Failed to fetch models:', error)

      // 连接错误处理（包括跨域检测）
      if (error.message && (error.message.includes('Failed to fetch') ||
          error.message.includes('Connection error'))) {
        const isCrossOriginError = this.detectCrossOriginError(error, baseURL)

        if (isCrossOriginError) {
          throw new APIError(`Cross-origin connection failed: ${error.message}`)
        } else {
          throw new APIError(`Connection failed: ${error.message}`)

View on GitHub (pinned to 3e677b1d9f)

Solutions

1. Check the backing server actually serves models: curl $BASE_URL/v1/models
2. If local, load/pull a model (ollama pull, vllm serve ...) before listing
3. If a project-scoped key is used, grant the project access to at least one model
4. Treat this as a non-fatal empty state in UI rather than crashing

Example fix

// before
const models = await adapter.getModelsAsync()
setModels(models)

// after
let models: TextModel[] = []
try { models = await adapter.getModelsAsync() }
catch (e) { if (!(e instanceof APIError && /empty model list/.test(e.message))) throw e }
setModels(models) // render empty state gracefully

const resp = await fetch(`${baseURL}/models`, { headers })
const body = await resp.json()
if (!Array.isArray(body?.data) || body.data.length === 0) warn('No models available on server')

null

try { models = await adapter.getModelsAsync() }
catch (e) {
  if (e instanceof APIError && /empty model list/.test(e.message)) models = []
  else throw e
}

When it happens

Trigger: Listing models from a server whose model list is empty, entries filtered out because model.id is missing/non-string, or a gateway exposing zero deployed models.

Common situations: Local gateways (vLLM/llama.cpp/Ollama bridges) with no models loaded yet, an OpenAI-compatible service where all entries lack the id field, freshly provisioned projects with no model access granted.

Related errors

• Unexpected API response format
• UNSUPPORTED_TEST_TYPE
• Unexpected API response format
• Cloudflare model search returned an unexpected response form
• Cross-origin connection failed: ${error.message}

AI-assisted analysis of linshenkx/prompt-optimizer@3e677b1d9f (2026-08-27). Data as JSON: /api/errors/917573cddeec10d7. Report an issue: GitHub.