Agents not working with Azure OpenAI ("No model spec selected")

[GPTLeonard]

What happened?

When I create a new Agent in LibreChat and select Azure OpenAI as the provider, I get the following error when trying to use the Agent:

Something went wrong. Here's the specific error message we encountered: No model spec selected

This happens regardless of which Azure OpenAI model I select (tested with gpt-4.1 and gpt-5-mini). I expect the Agent to use the selected Azure model normally, similar to how Anthropic or OpenAI works.

It looks like the Agent Builder does not pass a proper model spec for Azure OpenAI endpoints.

Version Information

LibreChat version: v0.8.0-rc2

Steps to Reproduce

1. Go to Agent Builder in LibreChat
2. Click "Create Agent"
3. Select Azure OpenAI as provider and choose any available model (e.g. gpt-4.1 or gpt-5-mini)
4. Try chatting with the Agent
5. Error occurs: "No model spec selected"

What browsers are you seeing the problem on?

Chrome

Relevant log output

Error message shown in UI:
"Something went wrong. Here's the specific error message we encountered: No model spec selected"

Screenshots

Code of Conduct

• I agree to follow this project's Code of Conduct

[danny-avila]

You have model specs enabled with some of the restrictive options.

Please review the config guide and share your config if you'd like help troubleshooting:
https://www.librechat.ai/docs/configuration/librechat_yaml/object_structure/model_specs

[GPTLeonard]

# LibreChat Configuration
version: 1.2.9
cache: true
fileStrategy: "azure_blob"

interface:
  endpointsMenu: false
  modelSelect: true
  parameters: false
  sidePanel: true
  presets: false
  prompts: true
  bookmarks: false
  multiConvo: false
  customWelcome: "Waar wil je vandaag samen aan werken?"
  agents: true
  runCode: false

# Speech configuratie
speech:
  speechTab:
    conversationMode: true
    advancedMode: false
    speechToText:
      engineSTT: "external"
      languageSTT: "Dutch (Netherlands)"
      autoTranscribeAudio: true
      decibelValue: -45
      autoSendText: 0
    textToSpeech:
      engineTTS: "browser"  # Browser TTS uitgeschakeld
      automaticPlayback: false

  stt:
    azureOpenAI:
      instanceName: '{example}'
      apiKey: '${STT_API_KEY}'
      deploymentName: 'whisper'
      apiVersion: '2024-06-01'

endpoints:
  azureOpenAI:
    titleConvo: true
    titleMethod: "completion"
    titleModel: "gpt-4.1-title"
    groups:
      - group: "{example}"
        apiKey: "${AZURE_OPENAI_API_KEY}"
        instanceName: "${AZURE_OPENAI_INSTANCE}"
        version: "preview"
        models:
          gpt-4.1:
            deploymentName: "${AZURE_GPT41_DEPLOYMENT}"
          gpt-5-mini:
            deploymentName: "${AZURE_GPT5_DEPLOYMENT}"

      - group: "{example}"
        apiKey: "${AZURE_OPENAI_API_KEY}"
        instanceName: "${AZURE_OPENAI_INSTANCE}"
        version: "2024-10-21"
        models:
          gpt-4.1-title:
            deploymentName: "${AZURE_GPT41_DEPLOYMENT}"

  # Custom endpoint voor websearch
  custom:
    - name: "websearch-api"
      apiKey: "${WEBSEARCH_API_KEY}"
      baseURL: "https://api.perplexity.ai/"
      models:
        default: ["sonar-pro", "sonar-deep-research"]
        fetch: false
      summarize: false
      forcePrompt: false
      dropParams: ["stop", "frequency_penalty"]
      modelDisplayLabel: "Web Search"

  agents:
    recursionLimit: 25
    maxRecursionLimit: 50
    disableBuilder: false
    allowedProviders:
      - azureOpenAI
    capabilities:
      - "file_search"
      - "actions"
      - "artifacts"

modelSpecs:
  enforce: true
  prioritize: false
  addedEndpoints:
    - agents
  list:
    - etc. All models.....

[jsiegmund]

@GPTLeonard you're issue might be with enforce: true under modelSpecs

[Priya-Navandar]

how to add tools to this agent?

[jsiegmund]

There's a button labeled 'Add Tools'