AI in your editor: why participants beat chat tabs

For two years, the default answer to 'how do I add AI to a code editor' was: add a chat tab. Cursor did it. Copilot Chat did it. Every startup building on top of VS Code did it. You get a sidebar with a message list, a text input at the bottom, and a model picker somewhere near the send button.

That pattern works. It also treats coding like a customer service problem. You visit the model, ask a question, wait for an answer, then return to your code.

The interesting question is what happens when the model stops being a place you visit and becomes a peer you can address.

Slack introduced it years ago. You do not visit @channel. You mention it. Discord does the same with bots. The name is a primary handle in the conversation you were already having.

VS Code shipped the same idea for its built in Chat panel. Any extension can register a participant, and users invoke it by typing @ followed by its handle. The extension picks up the message, produces a reply, and streams it back inline.

Suddenly the model is not a destination. It is available exactly where the thought happens, mixed in with your workspace context and prior turns.

Your code, your terminal, your file tree, and your open tabs already live in one namespace. A chat tab that lives beside them but cannot see them forces copy paste for every question. Every reply comes back as text you have to shepherd into the right file.

A participant lives inside the conversation about the code you are already looking at. VS Code sends the participant your prompt, the context from your workspace, and your prior chat history. The response can point at line ranges, apply edits, and reference symbols the user actually has open.

None of that is possible from a sidebar that watches the editor from the outside.

We shipped taskclan.taskclan-intelligence on the VS Code Marketplace this week. It registers @taskclan as a participant in the built in Chat panel. Ask it something plain and the router picks a T1 tier for you. Use a slash command and it routes to the tier that fits.

• /agent runs the full planning loop on T1 Max.
• /plan produces a plan with no code emission, still on T1 Max.
• /refactor rewrites what you have selected, on T1 Flow.
• /explain gives you the short breakdown on T1 Core.

Every reply footer tells you which tier actually ran, how many credits it cost, and your remaining balance. No hidden math, no black box billing.

Install the extension, paste an sk_t1_ key from platform.taskclan.com/api-keys, and type @taskclan hi in the Chat panel. That is the whole flow.

Marketplace listing: Taskclan Intelligence on the VS Code Marketplace. Docs walk you through the setup and every slash command at taskclan.com/docs/editors.

If you already use Continue.dev, the config path still works. Both routes end at the same T1 wallet.