6 replies

Openclaw in a ubuntu virtual Machine to use Ollama Model In Ubuntu Host

✅Solved

Hello!

I am trying to configue Openclaw in my virtual machine to use Ollama as the model provider.

Using postman I can do requestes from my virtual machine to Ollama in the host.

But when Openclaw tries using Ollama I see this logs in ollama errors:

⁨

Jan 31 20:41:03 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:03 | 404 |   68.912175ms | 192.168.122.233 | POST     "/v1/responses"
Jan 31 20:41:19 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:19 | 404 |   53.472129ms | 192.168.122.233 | POST     "/v1/responses"
Jan 31 20:41:21 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:21 | 404 |   50.361045ms | 192.168.122.233 | POST     "/v1/responses"
Jan 31 20:41:23 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:23 | 404 |   45.862158ms | 192.168.122.233 | POST     "/v1/responses"

Jan 31 20:41:03 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:03 | 404 |   68.912175ms | 192.168.122.233 | POST     "/v1/responses"
Jan 31 20:41:19 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:19 | 404 |   53.472129ms | 192.168.122.233 | POST     "/v1/responses"
Jan 31 20:41:21 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:21 | 404 |   50.361045ms | 192.168.122.233 | POST     "/v1/responses"
Jan 31 20:41:23 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:23 | 404 |   45.862158ms | 192.168.122.233 | POST     "/v1/responses"

⁩

Any Idea why this could be?

Solution

Option 2: Configure the correct endpoint
In your

~/.clawdbot/config.json5

~/.clawdbot/config.json5

, ensure you're specifying the Ollama provider correctly. You may need to explicitly set the API type:

{
  providers: {
    ollama: {
      baseUrl: "http://<host-ip>:11434/v1",
      api: "openai-completions",  // Explicitly use chat completions, not responses
      apiKey: "ollama"  // Required but ignored
    }
  }
}

{
  providers: {
    ollama: {
      baseUrl: "http://<host-ip>:11434/v1",
      api: "openai-completions",  // Explicitly use chat completions, not responses
      apiKey: "ollama"  // Required but ignored
    }
  }
}

Option 3: Use a reverse proxy (workaround)
Set up a simple proxy that forwards

/v1/responses

/v1/responses

requests to Ollama's

/v1/chat/completions

/v1/chat/completions

endpoint.

Jump to solution

FotC�

Friends of the Crustacean 🦞🤝•2mo ago•

6 replies

leviatan89

Openclaw in a ubuntu virtual Machine to use Ollama Model In Ubuntu Host

✅Solved

Jan 31 20:41:03 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:03 | 404 |   68.912175ms | 192.168.122.233 | POST     "/v1/responses"
Jan 31 20:41:19 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:19 | 404 |   53.472129ms | 192.168.122.233 | POST     "/v1/responses"
Jan 31 20:41:21 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:21 | 404 |   50.361045ms | 192.168.122.233 | POST     "/v1/responses"
Jan 31 20:41:23 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:23 | 404 |   45.862158ms | 192.168.122.233 | POST     "/v1/responses"

Jan 31 20:41:03 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:03 | 404 |   68.912175ms | 192.168.122.233 | POST     "/v1/responses"
Jan 31 20:41:19 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:19 | 404 |   53.472129ms | 192.168.122.233 | POST     "/v1/responses"
Jan 31 20:41:21 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:21 | 404 |   50.361045ms | 192.168.122.233 | POST     "/v1/responses"
Jan 31 20:41:23 leviatan89-System-Product-Name ollama[152440]: [GIN] 2026/01/31 - 20:41:23 | 404 |   45.862158ms | 192.168.122.233 | POST     "/v1/responses"

⁩

Any Idea why this could be?

Solution

Option 2: Configure the correct endpoint
In your

~/.clawdbot/config.json5

~/.clawdbot/config.json5

, ensure you're specifying the Ollama provider correctly. You may need to explicitly set the API type:

{
  providers: {
    ollama: {
      baseUrl: "http://<host-ip>:11434/v1",
      api: "openai-completions",  // Explicitly use chat completions, not responses
      apiKey: "ollama"  // Required but ignored
    }
  }
}

{
  providers: {
    ollama: {
      baseUrl: "http://<host-ip>:11434/v1",
      api: "openai-completions",  // Explicitly use chat completions, not responses
      apiKey: "ollama"  // Required but ignored
    }
  }
}

Option 3: Use a reverse proxy (workaround)
Set up a simple proxy that forwards

/v1/responses

/v1/responses

requests to Ollama's

/v1/chat/completions

/v1/chat/completions

endpoint.

Jump to solution

Openclaw in a ubuntu virtual Machine to use Ollama Model In Ubuntu Host

Openclaw in a ubuntu virtual Machine to use Ollama Model In Ubuntu Host

Similar Threads

Similar Threads

Similar Threads