GIDE 2.2.0: any GGUF, online models on your own key and Skill Files

GIDE 2.2.0 is the release where GIDE stops being a single-model editor. It runs several local model families, and for the first time it lets you bring your own key for online models. Skill Files teach GIDE your project’s conventions, and MCP servers extend what the agent can do.

At a glance

Auto-configured local models. Qwen, Ornith, Gemma and DeepSeek in GGUF.
Online models on your own key. Anthropic (Claude), OpenAI and Google Gemini.
Skill Files. Reusable guidance under .gide/skills.
MCP servers. Connect tools such as NVIDIA OpenShell to the chat agent.
Temperature and GPU offload. Per-model sliders with recommended values.

Auto-configured local models

What changed

GIDE now runs several local model families beyond the default: Qwen, Ornith, Gemma and DeepSeek in GGUF.

How it works

Pick the family and point GIDE at the model file. GIDE configures memory, context size and tool calling for that model automatically.

Online models on your own key

What changed

Bring your own API key to run Anthropic (Claude), OpenAI or Google Gemini directly inside GIDE.

Why it matters

Keys are stored locally in your system keychain and are only ever sent to the provider you choose. GIDE ships no key of its own. You decide which provider, if any, sees your code.

Skill Files

Write reusable skill files under .gide/skills in your project to teach GIDE your project’s conventions and procedures.

Skills activate deterministically: when you @mention them, or when a matching file is open. There’s no fuzzy guessing, so guidance is only applied when it fits.

MCP servers

Connect OpenShell to GIDE over MCP and let the chat agent call its tools directly. More generally, MCP extends GIDE beyond its built-in tools to whatever your connected MCP servers expose.

Model controls

Temperature

Every model now has a temperature slider in its settings, from precise to more creative. GIDE shows a recommended value per model and applies the change on your next message.

GPU offloading

A per-model offload slider controls how many model layers run on the GPU, with a recommendation based on your VRAM and RAM. Adjusting it reloads the model so the change takes effect right away.

A refreshed chat

The chat interface has a new toolbar, cleaner inline-diff controls and a polished reasoning view. The model’s thinking now streams and animates correctly for both local and online models.

Fixes

Extension compatibility. Fixed an issue that blocked 1.x extensions from installing on GIDE 2.x. Gallery assets and chat anchors now resolve reliably, so the extension marketplace works as expected.

For the short version of every change, see the GIDE changelog.