Ollama Cloud Provider Setup

Ollama Cloud serves as Ollama's hosted model API. The ollama-cloud provider reaches it directly at https://ollama.com through Ollama's native /api/chat API, requiring neither a local Ollama server nor a local Ollama app signed into cloud mode. Model references such as ollama-cloud/kimi-k2.6 are supported.

OpenClaw assigns ollama-cloud its own provider id, ensuring that cloud-only credentials, live catalog discovery, and model selection stay separate from a local ollama host. For local Ollama setups, hybrid cloud-plus-local routing, embeddings, and custom host configurations, refer to Ollama.

Setup

Begin by following Ollama's API key instructions, then execute:

openclaw onboard --auth-choice ollama-cloud

Alternatively, configure:

export OLLAMA_API_KEY="<your-ollama-cloud-api-key>" # pragma: allowlist secret

For non-interactive onboarding, the key can be passed directly:

openclaw onboard --auth-choice ollama-cloud --ollama-cloud-api-key "<key>"

The onboarding process sets ollama-cloud/minimax-m2.7 as the default model.

Defaults

Provider: ollama-cloud
Base URL: https://ollama.com
Env var: OLLAMA_API_KEY
API style: Ollama native /api/chat
Onboarding default model: ollama-cloud/minimax-m2.7

When to choose Ollama Cloud

Hosted Ollama models are what you need without running ollama serve locally.
The same native Ollama chat API shape OpenClaw uses for local Ollama is what you want, but aimed at https://ollama.com.
A straightforward cloud route for models already present in Ollama's hosted catalog is desired.
Local model pulls, local GPU control, or LAN-only inference are not required.

Choose Ollama when local-only or cloud-plus-local routing through a signed-in Ollama host is the goal. An OpenAI-compatible provider fits when /v1/chat/completions semantics or provider-specific OpenAI-style features are necessary.

Models

An API key is mandatory for this provider; without one, it remains inactive. With a key in place, OpenClaw pulls Ollama Cloud models live from the hosted catalog:

openclaw models list --provider ollama-cloud
openclaw models set ollama-cloud/kimi-k2.6

The live catalog contains hosted ids such as deepseek-v4-flash, glm-5.2, gpt-oss:20b, kimi-k3, and minimax-m3. Should live discovery come up empty, OpenClaw defaults to the bundled entries minimax-m2.7, minimax-m3, kimi-k3, glm-5.1, and glm-5.2. The retired kimi-k2.5 stays flagged as deprecated for existing exact references, though it no longer counts as a current hosted model.

Model ids correspond to cloud catalog ids, not local pull names. When a model name functions on a local Ollama host but is missing from the hosted catalog, turn to the ollama provider with that local host.

Live test

For Ollama Cloud API-key smoke tests, aim the Ollama live test at the hosted endpoint and pick a model from your current catalog:

export OLLAMA_API_KEY="<your-ollama-cloud-api-key>" # pragma: allowlist secret

OPENCLAW_LIVE_TEST=1 \
OPENCLAW_LIVE_OLLAMA=1 \
OPENCLAW_LIVE_OLLAMA_BASE_URL=https://ollama.com \
OPENCLAW_LIVE_OLLAMA_MODEL=kimi-k2.6 \
pnpm test:live -- extensions/ollama/ollama.live.test.ts

The cloud smoke covers text, native stream, and web search; set OPENCLAW_LIVE_OLLAMA_WEB_SEARCH=0 to bypass web search. Embeddings are skipped by default for https://ollama.com since Ollama Cloud API keys may not grant access to /api/embed; use OPENCLAW_LIVE_OLLAMA_EMBEDDINGS=1 to force them.

Troubleshooting

Ollama Cloud requires an API key / Set OLLAMA_API_KEY errors: supply a genuine cloud API key. The local ollama-local marker applies only to local or private Ollama hosts.
Unknown model errors: execute openclaw models list --provider ollama-cloud and replicate the hosted model id exactly.
Tool-call or raw JSON problems on custom Ollama hosts: verify you are not inadvertently using an OpenAI-compatible /v1 URL. Ollama routes should rely on the native base URL without any /v1 suffix.