> ## Documentation Index > Fetch the complete documentation index at: https://doc.onetoken.one/llms.txt > Use this file to discover all available pages before exploring further. # AI workbench ## AI workbench (OneToken account + API key) The AI workbench is a **web app** that uses your OneToken **LLM API key** to: * **Chat** (text, multi-turn, clear context) * **Image generation** (text-to-image; pick models and parameters) * **Video generation** (text-to-video; pick models and parameters) > OneToken: **API relay** (sign up, billing, API keys, model selection).\ > AI workbench: **browser UI** (calls with your key and shows results).\ > Open: **[https://ai.onetoken.one/](https://ai.onetoken.one/)** *** 11.png ## 1. Prerequisites * A OneToken account * A valid OneToken **API key** * Balance or quota (billing varies by model) *** ## 2. Register and get an API key 1. Sign up / log in on the OneToken site (follow on-page steps). 2. Open the console / dashboard. 3. Create and copy your **API key** * Prefer separate keys per use case for rotation and revocation. 4. (Optional) Confirm which models are enabled (chat, image, video). > Treat your API key like a password—do not share it. *** ## 3. Log in to the workbench 1. Open **[https://ai.onetoken.one/](https://ai.onetoken.one/)** 2. Sign in with your OneToken account (use one-click login if offered). 3. After login, check balance in the lower-left area. *** ## 4. Bind your OneToken API key After you register, a key is created and tied to the workbench—open the workbench and you can start. **Check that it works:** * Balance refresh succeeds. * You can send a chat and get a reply. *** ## 5. UI overview The left sidebar usually includes: * **AI chat** * **AI image** * **AI video** * **History**, **New chat**, etc. *** ## 6. Chat from scratch ### 6.1 New conversation * Click **New chat** * Choose **AI chat** ### 6.2 Pick a model Use the model dropdown (e.g. `gpt-5.2` in screenshots): * Stronger reasoning / long context: heavier chat models * Faster / cheaper: lighter models (per OneToken’s available list) ### 6.3 Send a message * **Enter** to send * **Ctrl/Cmd + Enter** for newline (follow on-page hints) ### 6.4 Common actions * **Clear context**: reset session memory * **New chat**: new thread * **History**: switch threads in the sidebar #### Example prompts * “Rewrite this paragraph in a more professional tone: …” * “Give me a 7-day fitness plan for beginners with equipment and diet sections.” * “From this requirements doc, output API definitions and field specs (JSON).” *** ## 7. Image generation (text-to-image) ### 7.1 Open image mode * Click **AI image** in the sidebar. ### 7.2 Choose an image model Pick from models OneToken exposes. ### 7.3 Write a prompt Include: * Subject (person, scene, object) * Style (realistic, anime, cyberpunk, watercolor, …) * Composition (close-up, full body, top-down, wide) * Light and mood (soft light, backlight, neon night) * Quality hints (high detail, 8k, cinematic) Optional negative prompt, e.g.: * “low quality, blurry, bad hands, extra fingers” ### 7.4 Parameters and generate Typical controls: * Aspect ratio (1:1, 3:4, 16:9, …) * Resolution * Batch count * Seed Click **Generate** and wait for results. #### Example prompt > “An orange cat in an astronaut helmet floating inside a space capsule, blue Earth in the window, cinematic lighting, ultra-detailed, photorealistic.” *** ## 8. Video generation (text-to-video) ### 8.1 Open video mode * Click **AI video**. ### 8.2 Choose a video model Pick from models OneToken exposes. ### 8.3 Prompt tips Include: * Scene and subject * Camera motion (push in, pull out, orbit, follow, pan) * Duration and pacing (if supported) * Style (realistic, animated, film-like) ### 8.4 Parameters and generate Common options (when available): * Duration (seconds) * Resolution (720p / 1080p) * Frame rate * Aspect ratio (16:9 / 9:16) Video jobs take longer—please wait. #### Example prompt > “Rainy city street at night, neon reflections on wet pavement, a person in a trench coat walks right to left, slow tracking shot, shallow depth of field, realistic.” *** ## 9. Billing * The workbench **issues requests and shows output**. * **Charges** follow OneToken rules for the model and usage (tokens, image/video jobs, resolution, duration, etc.). Tips: * Start with cheaper models for tests. * For images/video, try low resolution / short clips first. *** ## 10. FAQ ### Q1: “Key not configured / no permission / 401” * Confirm the key is active and not disabled. * Refresh the page and **refresh balance**. ### Q2: Failures or timeouts? * Image/video are slower—retry or wait. * Switch model or lower resolution, duration, or batch size. * Check OneToken status and balance. ### Q3: Why does “clear context” change behavior? * The model no longer sees prior messages—useful for a fresh topic. *** ## 11. Security * Do not paste keys in chats, tickets, or public repos. * Rotate keys periodically; revoke leaked keys immediately. * For teams, use per-member or per-project keys. *** ## 12. One-minute checklist 1. Sign up / log in to OneToken 2. Create and copy an API key 3. Open [https://ai.onetoken.one/](https://ai.onetoken.one/) and sign in 4. In **admin / settings**, paste the key if prompted and save 5. In **AI chat**, pick a model and send: “Hi, write a short product blurb.” 6. Try **AI image** and **AI video** ***