Documentation

1. Create an account.
2. New accounts are placed on the Free plan and receive starter credits.
3. Open Chat to start a conversation, or create an API key in Dashboard → API keys to integrate GLAVYN AI into your own software.

• Streaming: replies appear as they are generated. Press Stop to end a reply early; you are charged only for what was generated.
• Edit and regenerate: hover a message to copy it, edit a previous prompt (later messages are replaced), or regenerate the last reply.
• Files: attach images (PNG, JPEG, WebP, HEIC), PDFs, or text and code files. Files are private to your account.
• Models: choose a model from the picker at the top of a conversation. The choice is remembered per conversation.
• History: conversations are saved automatically. Use search, rename, pin, and delete from the sidebar.
• Press Enter to send and Shift + Enter for a new line.

GLAVYN AI remembers the conversation you are in. It can also keep long-term memories: short facts and preferences you share (for example, “Prefers TypeScript”) that make future answers more relevant.

• Memories are extracted in the background after a reply and are visible in Dashboard → Memory.
• You can delete individual memories, clear all of them, turn memory off, or require approval before new memories are saved.
• Memory never stores passwords, API keys, card numbers, ID numbers, or contact details, and it is never shared with other accounts.

API keys look like gl_live_… and are shown only once, when created. GLAVYN AI stores only a keyed hash, so a lost key cannot be recovered; revoke it and create a new one. Keys can have an expiry date and their own request-per-minute limit.

See the API reference for endpoints and examples.

The GLAVYN AI extension brings the chat into Visual Studio Code and editors built on it (Cursor, Windsurf, VSCodium, Kiro), in the side bar on the right. It sends the file you have open or your selection with each question, and writes the code from replies straight into your files.

• Changes: each file is written as soon as its code is complete, and the chat shows a card with a diff and Undo. Hidden files, files with unsaved edits, and changes that would remove most of a file wait for Apply. To approve every change yourself, switch Auto to Review under the message box (setting glavyn.applyChanges).
• Sign in: click Sign in in the GLAVYN AI panel, approve on this site, and the browser returns you to the editor. If it cannot, copy the code shown and run GLAVYN AI: Enter Sign-In Code.
• Keys: each signed-in editor gets its own API key, named after the editor and the computer, in Dashboard → API keys. It counts towards your plan's key limit; signing out revokes it.
• Usage: requests from the extension use your plan's versions, limits and credits, like the API.

• Credits: each request is charged from your credit balance at the model's per-token price. Requests that your balance cannot cover are rejected with HTTP 402.
• Plan limits: plans set requests per minute, optional daily request and monthly token limits, the maximum output length, and which models you can use.
• Buying: upgrade your plan (prepaid for 1, 3, 6, or 12 months) or buy extra credits in Dashboard → Billing. Payments are processed securely by Razorpay (cards, UPI, net banking, and wallets).
• Spending limit: set your own monthly spending cap in Dashboard → Billing.
• Usage for every request (tokens, cost, latency) is listed in Dashboard → Usage.

• glavyn-2.8: GLAVYN Ultra (default). The most capable model, for complex questions, coding, analysis, images and PDFs.
• glavyn-2.5: GLAVYN Core. Fast, balanced model for everyday chat, writing, coding help and documents.
• glavyn-2.0: GLAVYN Flash. Lightweight model for quick questions, summaries and short drafts.
• test: Test.

Each model is a version of GLAVYN AI. Which versions you can use depends on your plan; the pricing page lists them for every plan, with current prices.

Sessions use secure, HTTP-only cookies; passwords are hashed with Argon2id; and every account's conversations, files, memories, keys, and usage are isolated. Read the privacy policy for how data is processed.