---
title: "Using Ollama Cloud models with OpenClaw at no cost"
description: "How I configured OpenClaw to use MiniMax, Kimi, and GLM cloud models through a local Ollama server, without a credit card or subscription."
date: "2026-03-03T00:00:00.000Z"
author: "Carlos Garavito"
tags: ["ai", "openclaw", "ollama", "free", "budget"]
canonical_url: "https://cgaravito.dev/en/blog/free-ai-ollama-cloud-openclaw"
last_updated: "2026-03-03T00:00:00.000Z"
locale: "en"
---

Ollama Cloud offered MiniMax M2.5, Kimi K2.5, and GLM-5 at no cost when I set up OpenClaw, without asking for a credit card or subscription. OpenClaw talks to a local Ollama server, and Ollama routes the request to the selected cloud model.

The setup took about five minutes.

The setup used these model identifiers:

- `minimax-m2.5:cloud` for coding and productivity
- `kimi-k2.5:cloud` for multimodal reasoning
- `glm-5:cloud` for reasoning and code generation

Ollama listed other models alongside these three. The available catalog, price, and usage limits can change, so I would check Ollama before treating those conditions as current.

## Configure the local path

Install OpenClaw without running onboarding.

```bash
curl -fsSL https://openclaw.ai/install.sh | bash -s -- --no-onboard
```

Install Ollama.

```bash
curl -fsSL https://ollama.com/install.sh | sh
```

Start the local Ollama server.

```bash
ollama serve &
```

Open the OpenClaw configuration flow through Ollama.

```bash
ollama launch openclaw --config
```

Select one of the cloud model identifiers from the list:

- `minimax-m2.5:cloud`
- `kimi-k2.5:cloud`
- `glm-5:cloud`

Start the OpenClaw gateway.

```bash
openclaw gateway &
```

Then open the terminal interface.

```bash
openclaw tui
```

## Check the generated configuration

The relevant part of my `openclaw.json` looked like this:

```json
{
  "models": {
    "providers": {
      "ollama": {
        "baseUrl": "http://127.0.0.1:11434",
        "apiKey": "ollama-local",
        "api": "ollama",
        "models": [
          {
            "id": "minimax-m2.5:cloud",
            "name": "minimax-m2.5:cloud",
            "reasoning": true,
            "contextWindow": 204800,
            "maxTokens": 128000
          }
        ]
      }
    }
  },
  "agents": {
    "defaults": {
      "model": {
        "primary": "ollama/minimax-m2.5:cloud"
      }
    }
  }
}
```

OpenClaw connects to `http://127.0.0.1:11434` and uses `"ollama-local"` as the API key. The local endpoint is the provider interface even though the selected model runs through Ollama Cloud.

| Model | Best For |
|-------|----------|
| `minimax-m2.5:cloud` | Coding, productivity |
| `kimi-k2.5:cloud` | Multimodal, reasoning |
| `glm-5:cloud` | Code generation |

## Fix the two failures I hit

If the cloud models do not appear, sign in and pull the model explicitly.

```bash
ollama signin
ollama pull minimax-m2.5:cloud
```

If the gateway does not start, restart it.

```bash
openclaw gateway restart
```

This configuration kept the OpenClaw side small, one local provider endpoint, one model identifier, and the generated `ollama-local` credential. The part I would recheck before using it again is the Ollama Cloud catalog and its current pricing.