Documentation

My AI Weapon is in invite-only early access. You'll need an invite code to create an account.

Once you have a code, sign up at myaiweapon.com/sign-up, then log in from any modern browser. Your workspace is ready immediately — no installer, no environment setup.

Open the chat and describe your first task in plain language. The agent plans the work, executes in your cloud environment, and reports progress back in the conversation.

My AI Weapon is chat-first. The conversation thread is your primary workspace — not a sidebar inside a code editor.

When you send a message, the agent breaks the task into steps: planning, writing files, running commands, calling APIs. Each step appears in the thread so you always know what's happening.

Results — code changes, previews, summaries — stay in the same conversation. No hunting through file trees or terminal tabs.

My AI Weapon uses a subscription + credits model. Your plan sets a monthly credit pool; usage deducts from that balance.

Credits cover both AI model calls and paid data APIs (web search, SEO data, news, and more). One balance, one bill — no separate API keys.

When credits run out, AI and data queries pause until your next billing cycle, you upgrade your plan, or you purchase a one-time credit pack.

Every account gets one persistent cloud environment. It's bound to your account, not a single device — log in from any browser and pick up where you left off.

Environments sleep after 30 minutes of idle time to save resources. Storage and projects persist. Waking takes under 5 seconds.

Higher plans include more compute and storage for heavier workloads.

By default, the model aggregator runs in Auto mode. It classifies each task and routes to the right tier — you don't pick models manually.

Simple tasks (summaries, small edits, short Q&A) use lightweight, efficient models. Code generation, multi-file changes, and agent steps use mid-tier models. Deep reasoning tasks can escalate to heavy models.

You can override Auto when you want a specific tier or model. Heavy models consume significantly more credits.

Your workspace uses opinionated file rules so you always know where things go. Projects live under ~/workspace/ with predictable structure.

Every new project gets a docker-compose.yml by default. Docker and Compose v2 are pre-installed — running apps in containers is the standard path, not a weekend project.

Platform logs go to ~/workspace/.logs/ (retained 7 days). Temporary files in ~/workspace/.tmp/ are cleaned daily.