Chat with T1 — Taskclan's AI tier system — inside VS Code and Cursor. Adds @taskclan to the built-in chat, with /agent, /plan, /refactor, and /explain slash commands. One key, four tiers auto-routed per prompt, metered against your Taskclan wallet.
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Chat with T1 — Taskclan's AI tier system — inside VS Code and Cursor. Adds @taskclan to the built-in chat, with slash commands that route to the right tier automatically.
Command palette → Taskclan: Set API Key → paste your sk_t1_… key.
Open Chat (Ctrl/Cmd + Alt + I) and type @taskclan help me refactor this.
Slash commands
Command
Runs on
What it does
@taskclan <prompt>
T1 Auto (or your default)
General chat, coding help, questions. Router picks the tier per prompt.
@taskclan /agent <prompt>
T1 Max
Agent-mode reply — plan, execute, ship.
@taskclan /plan <prompt>
T1 Max
Plan the change without touching code. Extended thinking.
@taskclan /refactor <prompt>
T1 Flow
Refactor the selection or referenced file.
@taskclan /explain <prompt>
T1 Core
Explain the selection, file, or symbol.
Settings
taskclan.apiBase — base URL for the OpenAI-compat endpoint. Defaults to https://engine.taskclan.com/api/openai/v1. Override for staging.
taskclan.defaultTier — which tier @taskclan uses without a slash command. Defaults to t1-auto (router picks per prompt).
taskclan.maxTokens — cap on output tokens per reply. Higher = longer answers but more credits.
What runs where
T1 Auto (default) — router reads your prompt and picks the tier.
T1 Core — fast + cheap, Haiku-class.
T1 Flow — balanced default, Sonnet-class.
T1 Max — deep reasoning, extended thinking, PDFs, prompt caching.
Every reply shows which tier actually ran, how many credits it cost, and your remaining balance.
Billing
Metered against your Taskclan wallet. Top up at platform.taskclan.com/billing. If you run out mid-session, the reply ends with a top-up link — no surprise overages.
