Models, without the guesswork
Free cloud.
Or your own computer.
OliveCode is a terminal AI coding agent for macOS, Windows and Linux. Connect a free cloud provider account or run a downloaded Ollama model locally.
How free cloud models work
Connect your own account through Cloud in OliveCode. The app shows eligible free models from connected providers. Availability, model names and provider offers can change; the in-app list is the place to check what is available to you now.
OliveCode does not sell a subscription or model credits. A provider may still require an account, set a request quota, restrict availability or change its free offer. OliveCode does not include paid-only models or promise free access to every leading proprietary model.
What happens when a provider reaches its limit?
OliveCode tries another eligible, available free model from your connected providers. This helps you continue when an alternative is available. If all eligible cloud options are unavailable, you may need to wait for a provider reset, connect another free provider or choose a local model.
Automatic switching is a fallback, not an unlimited cloud service. Different models can produce different results. Review the model shown for each completed answer and check generated work before using it.
Local coding models through Ollama
Open Local or type /local. OliveCode checks the computer’s hardware and estimates a model memory budget, then offers model choices suited to different sizes and capabilities. The estimate is guidance, not a guarantee: other apps, model context and runtime memory use affect what will fit.
Models are downloaded through Ollama and use disk space. A downloaded local model runs inference on your computer without a cloud-model request quota. Your RAM, storage and processor set practical limits. Downloads, provider connections, browser tasks and other online tools can still need an internet connection.
Use Delete models in the Local menu to manage installed models. Ollama’s default model folder is ~/.ollama/models on macOS, /usr/share/ollama/.ollama/models for the standard Linux service, and %USERPROFILE%\.ollama\models on Windows; an OLLAMA_MODELS override or a different Ollama service account can change it. See Ollama’s storage documentation.
Which option should I start with?
| Your priority | Start here |
|---|---|
| Quick start without a model download | A connected free cloud provider |
| Local inference without a cloud request quota | An Ollama model that fits your hardware |
| Limited laptop memory or disk space | Cloud, or a smaller local model |
| More perspectives on a task | Enable Assist with an available model |
There is no single best model for every task or computer. Test your actual coding task, inspect the changes, and run the project’s checks. A larger model or an extra review is not proof of correctness.