'Agent Not Found' Error

[Ria-c3alabs]

What happened?

When attempting to use custom contract agents configured in librechat.yaml, the system consistently returns an "Agent not found" error. The error appears in the UI as "An error occurred while processing your request. Please contact the Admin." In the logs, it shows "error: [handleAbortError] AI response error; aborting request: Agent not found".

Environment Details

LibreChat Version: Latest (as of April 2025)
Deployment Method: Docker
Configuration Method: Using librechat.yaml for agent definitions
Provider: OpenAI
Browser: Chrome (latest)

Error Messages

LibreChat | 2025-04-26 12:37:59 error: [handleAbortError] AI response error; aborting request: Agent not found
LibreChat | 2025-04-26 12:37:59 warn: Invalid conversation ID: undefined
LibreChat | 2025-04-26 12:37:59 info: ---`saveMessage` context: undefined
LibreChat | 2025-04-26 12:37:59 info: ---Invalid conversation ID Params: {
LibreChat |   "messageId": "d4429f7a-9c5a-478c-9442-f33f18427115",
LibreChat |   "parentMessageId": "00000000-0000-0000-0000-000000000000",
LibreChat |   "unfinished": false,
LibreChat |   "error": true,
LibreChat |   "final": true,
LibreChat |   "text": "An error occurred while processing your request. Please contact the Admin.",
LibreChat |   "isCreatedByUser": false,
LibreChat |   "spec": "contract-search-agent",
LibreChat |   "iconURL": "https://cdn.iconscout.com/icon/premium/png-256-thumb/contract-8597389-6953248.png",
LibreChat |   "agent_id": "contract_search_agent",
LibreChat |   "user": "680cc75d0da1a0c630ec1846"
LibreChat | }

Librechat.yaml config:

• added modelSpecs:

modelSpecs:
  enforce: true
  prioritize: true
  list:
    - name: "contract-search-agent"
      label: "Contract Search Assistant"
      default: true
      description: "Specialized agent for searching and analyzing contracts"
      iconURL: "https://cdn.iconscout.com/icon/premium/png-256-thumb/contract-8597389-6953248.png"
      showIconInMenu: true
      showIconInHeader: true
      preset:
        endpoint: "agents"
        agent_id: "contract_search_agent"  # Add "agent:" prefix here

    - name: "contract-generator-agent"
      label: "Contract Generator"
      description: "Specialized agent for generating legal contracts"
      iconURL: "https://cdn.iconscout.com/icon/premium/png-256-thumb/contract-11759188-9601742.png"
      showIconInMenu: true
      showIconInHeader: true
      preset:
        endpoint: "agents"
        agent_id: "contract_generator_agent"

• interface config:

endpointsMenu: true
  modelSelect: true
  parameters: true
  sidePanel: false
  presets: true
  prompts: true
  bookmarks: true
  multiConvo: true
  agents: true
  showAgentIcons: true
  showModelIcons: true
  agentsMenu: true
  showAgentsInModelSelect: true
  agentBuilder: true

• endpoint config:

agents:
    enabled: true
    defaults:
      - id: "contract_search_agent"
        name: "Contract Search Assistant"
        allowedProviders: ["openai"]
        description: "Specialized agent for searching and analyzing contracts"
        instructions: |
          You are a specialized Contract Search and Analysis agent with expertise in legal document analysis. Your purpose is to help users search, analyze, and extract information from contracts with precision and accuracy.
        provider: "openai"
        model: "gpt-3.5-turbo"
        model_parameters:
          temperature: 0.2
          top_p: 1
        tools: ["file_search"]
        capabilities: ["file_search", "tools"]

      - id: "contract_generator_agent"
        name: "Contract Generator"
        allowedProviders: ["openai"]
        description: "Specialized agent for generating legal contracts"
        instructions: |
          You are a specialized Contract Generation agent with expertise in legal document creation. Your purpose is to help users create new contracts based on their specific requirements and past contract templates.
        provider: "openai"
        model: "gpt-3.5-turbo"
        model_parameters:
          temperature: 0.5
          top_p: 1
        tools: ["file_search"]
        capabilities: ["file_search", "tools"]

Kindly resolve this.

Version Information

ghcr.io/danny-avila/librechat-dev latest 74b3b760963f 9 hours ago 1.62GB
ghcr.io/danny-avila/librechat-rag-api-dev-lite latest 566585698235 2 weeks ago 1.85GB

Steps to Reproduce

1. Configure custom agents in librechat.yaml as described above
2. Restart LibreChat using docker
3. Select one of the custom agents from the model selection dropdown
4. Enter a prompt and submit
5. Observe the "Agent not found" error

What browsers are you seeing the problem on?

No response

Relevant log output

LibreChat | 2025-04-26 12:37:59 error: [handleAbortError] AI response error; aborting request: Agent not found
LibreChat | 2025-04-26 12:37:59 warn: Invalid conversation ID: undefined
LibreChat | 2025-04-26 12:37:59 info: ---`saveMessage` context: undefined
LibreChat | 2025-04-26 12:37:59 info: ---Invalid conversation ID Params: {
LibreChat |   "messageId": "d4429f7a-9c5a-478c-9442-f33f18427115",
LibreChat |   "parentMessageId": "00000000-0000-0000-0000-000000000000",
LibreChat |   "unfinished": false,
LibreChat |   "error": true,
LibreChat |   "final": true,
LibreChat |   "text": "An error occurred while processing your request. Please contact the Admin.",
LibreChat |   "isCreatedByUser": false,
LibreChat |   "spec": "contract-search-agent",
LibreChat |   "iconURL": "https://cdn.iconscout.com/icon/premium/png-256-thumb/contract-8597389-6953248.png",
LibreChat |   "agent_id": "contract_search_agent",
LibreChat |   "user": "680cc75d0da1a0c630ec1846"
LibreChat | }

Screenshots

No response

Code of Conduct

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

[danny-avila]

Are you using the actual agent ID below the name?

[Ria-c3alabs]

Yes, both the agent_ids here are correctly set in modelSpecs and also match the ID in my agent definition.

[danny-avila]

There are some issues with your config. Sadly I’m on my phone and can’t run it through an AI but I would ask you to review the documentation for both agents and model specs.

https://www.librechat.ai/docs/configuration/librechat_yaml/object_structure/agents

https://www.librechat.ai/docs/configuration/librechat_yaml/object_structure/model_specs

[Ria-c3alabs]

@danny-avila I reviewed the entire docs, and I am still unable to find the issue. I have also added allowedProviders key but still the same issue. Can you please guide me through the issue?

[Ria-c3alabs]

@danny-avila, kindly help me with this, as I have been stuck for a long time now.

[danny-avila]

First, the options you specified for the agents config don't even exist, so that could be creating issues. Reviewing the documentation would show this to you.

Here's the relevant parts of my librechat.yaml for you to use as a reference:

endpoints:
  agents:
    disableBuilder: false
interface:
  agents: true
  prompts: true
  parameters: true
  presets: true
  runCode: true
  modelSelect: false
  bookmarks: true
  multiConvo: true
modelSpecs:
  enforce: false
  prioritize: true
  list:
    - name: "Claude Coder"
      label: "Claude Coder"
      iconURL: https://cdn1.cloud.e-infra.cz/logos/meta-llama.png
      preset:
        endpoint: "agents"
        agent_id: "agent_someAgentId"

No issues

[iamronr]

I'm getting the same error. I'm trying to use Qwen 3 and add Google search via the agent.

modelSpecs

list:
    - name: "Qwen3 Agent"
      label: "Qwen3-8B-Q8 Agent"
      default: false
      description: "Qwen3 w/Google Search"
      iconURL: "https://huggingface.co/spaces/Qwen/Qwen3-Demo/resolve/500a858287c9d594643270b80f1e28f6e9b89c9c/assets/qwen.png"
      preset:
        endpoint: "agents"
        agent_id: "qwen3agent"

agents

enabled: true
    defaults:
      - id: "qwen3agent"
        name: "Qwen3 Agent"
        allowedProviders: ["google"]
        description: "Qwen3 with Google Search access"
        instructions: |
        provider: "Ollama - rGPT"
        model: "hf.co/unsloth/Qwen3-8B-GGUF:Q8_0"
        model_parameters:
          temperature: 0.2
          top_p: 1
        tools: ["tools"]
        capabilities: ["tools"]

Not sure what I'm missing...

[iamronr]

Nvm, got it. Forgot to use the actual agent ID...