Runs onThe places the vendor's own pages say the product runs: your machine, a container environment you run, or a cloud service the vendor runs. It is not a limit.
Models runWhere the vendor's own pages say the product's models run, not a limit: your machine, a container environment you run, the model company's cloud (servers of the company that makes or serves the model), or a cloud provider's model service. For a cloud, read the data and compliance lines of the company running it, linked by its name.
Several providers in one sessionOne session can use models from different providers, for example one agent on one maker's model and another on a second maker's. Switching the single model for the whole session does not count on its own. A tool that is itself a gateway or library routing to the providers its page names counts.
Several endpoints built inThe tool itself can hold more than one provider or address, so nothing has to sit in front of it to reach different providers.
Local and cloud togetherA local model (on your machine or in a container environment you run) and a cloud model (in its model company's cloud or a cloud provider's model service) can both be used in the same session or setup.
Use a local model: How to set it up(www.promptfoo.dev page, read 2026-10-19) Settings the guide names: OLLAMA_BASE_URL, OLLAMA_API_KEY, providers.
Use a local and a cloud model together: How to set it up(www.promptfoo.dev page, read 2026-10-22) Settings the guide names: providers.
Hold several endpoints in the product: How to set it up(www.promptfoo.dev page, read 2026-10-22) Settings the guide names: providers, OPENAI_API_KEY, ANTHROPIC_API_KEY.
Each item below is our own view of how the product compares or groups with other products in the catalog, or of a limit worth knowing about. It goes beyond what the maker's pages cited here say; no maker has said it or reviewed it, and it cannot be confirmed on a vendor's page. A second AI model read each item on 2026-10-24. Terms
It is closer to what a CI pipeline would call a test runner than to a dashboard-first observability platform. (as of 2026-09-24)
Data and compliance
Compliance: Set by whoever runs itThe answer belongs to whoever runs the software, because open-source software has no maker running it for you., for their own deployment.
Data: Set by whoever runs itThe answer belongs to whoever runs the software, because open-source software has no maker running it for you., with telemetry not checked yetWe have not read a page of the right kind for this question yet..
Model and where it runs: Run by its operator.
Details: What the project's own pages state, not a compliance determination.
What the maker's own pages said on the dates shown, not a compliance determination.
GitHub page: promptfoo(the product's own site, from our list of products; not read for these lines)
The readings behind this page
Our summary of the vendor's pages as of the dates shown on each line (2026-10-20 to 2026-10-24); confirm on the vendor's page.
Local models
promptfoo's Ollama page says an Ollama model can act as the local grader for test assertions that require a language model to evaluate them. (www.promptfoo.dev page, read 2026-10-20)
One sentence naming both a local and a cloud path
promptfoo allows comparing models side-by-side, including OpenAI, Anthropic, Azure, Bedrock, and Ollama. (github.com page, read 2026-10-20)
What a local model must meet
promptfoo's Ollama page says context windows are small for embedding models, with all-minilm defaulting to 256 tokens and topping out at 512. (www.promptfoo.dev page, read 2026-10-23)
promptfoo's Ollama page says small models are handy for trying out a config without a big download, naming qwen3:0.6b, about 500MB, which handles tools and reasoning. (www.promptfoo.dev page, read 2026-10-23)
promptfoo's Transformers provider page says that on limited RAM, running promptfoo eval with -j 1 runs the evals one at a time. (www.promptfoo.dev page, read 2026-10-23)
Plan, policy and billing
promptfoo's Anthropic page says requests sent with a Claude Code login are expected to draw on your Claude subscription as Claude Code CLI calls do, and points readers to Anthropic's own documentation on current billing. (www.promptfoo.dev page, read 2026-10-23)
Anthropic's legal and compliance page says Anthropic does not allow third-party developers to offer a Claude.ai login in their own apps, or to send their users' requests through credentials from a Free, Pro or Max plan. (code.claude.com page, read 2026-10-23)
Anthropic's legal and compliance page says OAuth authentication is intended only for people who bought a Claude plan, whether Free, Pro, Max, Team or Enterprise, and is meant for everyday use of Claude Code and Anthropic's other native applications. (code.claude.com page, read 2026-10-23)
Anthropic's legal and compliance page says developers whose products or services use Claude, those built on the Agent SDK among them, should use API keys, whether from Claude Console or from a cloud provider Anthropic supports. (code.claude.com page, read 2026-10-23)
Anthropic's legal and compliance page adds that these limits do not stop an end user from logging in to an unmodified copy of Claude Code with that user's own Claude plan, even on a platform that hosts Claude Code. (code.claude.com page, read 2026-10-23)
A separate model for planning and for doing
promptfoo's prompt optimization page says the provider you select is the one being optimized, and promptfoo's default suggestions provider, not that selected provider, writes the candidate prompt rewrites. (www.promptfoo.dev page, read 2026-10-23)
About the product
promptfoo's README says its LLM evals run entirely locally and that prompts never leave the user's machine. (github.com page, read 2026-10-24)
Our reading of vendors' pages, as of the dates shown. Confirm on the vendor's page before relying on any line that cites it. To report an error write to report@crinaro.ai