Inco for GitHub Copilot Chat

Use Inco AI models directly from the GitHub Copilot Chat model picker in Visual Studio Code.

This extension is a native VS Code LanguageModelChatProvider. It validates your Inco API key, discovers the models available to your workspace, and streams OpenAI-compatible chat completions directly from https://api.inco.ai/v1 into Copilot Chat.

Highlights

• Direct Inco integration without a local proxy
• API keys owned by VS Code native provider configuration
• Multiple isolated entries, identified by their group names in Manage Language Models
• Live /v1/models discovery with durable per-credential catalog snapshots
• Streaming text, reasoning output, images where advertised, token usage, and parallel function-tool calls
• Per-route reasoning and tool capabilities based on live probes
• Selectable context budgets, cancellation, and total/idle stream timeouts
• Bounded retries for pre-stream transient failures only
• Device-local token usage and estimated spend, with an authoritative dashboard link
• An optional Inco-served-only filter that hides relayed and unclassified models

Quick start

1. Install the extension. You need VS Code 1.125 or newer and GitHub Copilot Chat.
2. Create an API key in Inco Platform, or run Inco: Open API Keys.
3. Open Copilot Chat, select Manage Models, add an Inco provider entry, name it, and enter the key.
4. Choose a model available to that workspace.

To use more than one key, add another Inco entry with a different name. The entry name is its identity: VS Code namespaces each entry's models per group, stores the key securely, and offers key updates on the entry. Deleting an entry removes its VS Code-owned credential. Commands and local usage use the most recently provisioned entry; chat requests use their selected entry's exact key.

Use Inco: Manage Connection to open native provider configuration, Inco: Test Inference to verify a real response, and Inco: Show Local Usage to inspect device-local activity. The extension does not load .env during normal use.

Routing and data settings

Set incoCopilot.incoServedOnly to true to show and allow only verified Inco-served models. The filter also applies to cached and fallback catalogs and blocks requests from previously selected hidden models. Unknown operators stay hidden until classified. Some standard-tier models are Inco-served; a :fast suffix alone does not establish routing.

This routing filter does not guarantee zero data retention. Inco's dashboard Store & Improve, Training, and Allow relayed models controls determine storage, training, and relay access. See the current Inco privacy policy and security model.

Documentation

• Setup, settings, and troubleshooting
• Model capability differences, reasoning controls, and pricing
• API key and security model
• Development and releases

Related projects

• Codex Bridge for Copilot Chat — Use OpenAI models in Copilot Chat with ChatGPT authentication.
• Grok for GitHub Copilot Chat — Use xAI Grok models directly from the model picker.
• Inception for GitHub Copilot Chat — Use Inception Mercury models directly from the model picker.
• Ollama Cloud for GitHub Copilot Chat — Use Ollama Cloud models with native thinking and tool support.
• OpenCode for GitHub Copilot Chat — Use OpenCode Console and Go models from the model picker.
• Orvix for GitHub Copilot Chat — Use Orvix managed and BYOK models from the model picker.
• Poolside for GitHub Copilot Chat — Use hosted Poolside coding models from the model picker.

Unofficial project; not affiliated with Inco AI, GitHub, or Microsoft. Inco usage limits and charges still apply. Licensed under MIT.