BlueGPT
InspireLabs · Other
BlueGPT, now renamed VelvetYuzu, is an all-in-one AI platform that bundles 40+ chat, image, and video models with AI agents and PDF analysis under one credit plan.
About BlueGPT
What Is BlueGPT
BlueGPT is an all-in-one AI platform built by InspireLabs, a French company. It works as an AI model aggregator: one login, one credit balance, and more than 40 models from OpenAI, Anthropic, Google, Mistral, Meta, and Perplexity in a single chat window. Image and video generators, custom AI agents, and document analysis sit on top of that. The company renamed the product VelvetYuzu in 2026. Type bluegpt.app into a browser today and you land on velvetyuzu.ai. Most reviews and search results still carry the old name.
The problem it solves is subscription sprawl. A writer might keep a chat plan for drafting, an image tool for thumbnails, and a separate PDF assistant for research, and none of them share history. BlueGPT charges one credit balance that covers the whole workspace. Text, images, video, web search, and Deep Research all draw from that same pool, and folders keep each client's work apart.
The limits show up under load. Premium models burn AI credits faster than small ones, so a heavy afternoon of image generation can empty a month's allowance. Video is the tightest cap, and the free account is closer to a demo than a working setup. There's no public API and no native mobile app yet, so the product lives in a browser tab.
Getting Started
- Open bluegpt.app and sign up with an email address or a Google account. The domain redirects to VelvetYuzu, and the free account doesn't ask for a card.
- Confirm your email, then pick a paid plan if you need a larger monthly allowance. Starter runs €19 a month, or €15.83 when you pay yearly.
- Choose a model from the input bar, or open AI Agents and start from a preset such as the Perfect Prompt Generator or the SEO Rater.
- Attach a PDF or an image, or switch on web search, then send your prompt. Every run spends credits from the same monthly balance.
- Save the result to a folder, download the images or clips you generated, or share the chat with a colleague.
Product Information
A quick look at BlueGPT's pricing, supported platforms, and performance.
Best for
The users, tasks, and scenarios where this tool fits best.
Users
- Content creators and social media managers
- Marketing teams and small agencies
- Students and analysts who read long documents
Tasks
- Drafting, rewriting, and translating copy
- Generating images and short videos for campaigns
- Summarizing reports and pulling out figures
Scenarios
- Filling a week of social posts in one sitting
- Testing two models on the same brief
- Building a repeatable agent for a recurring job
Key features
40+ Models in One Chat
The workspace gathers text models from OpenAI, Anthropic, Google, Mistral, Meta, Microsoft, Cohere, and Perplexity, plus image engines such as DALL·E 3, Imagen 3, FLUX, Ideogram, and Stable Diffusion 3.5. You switch models inside a thread, so comparing two answers takes one click instead of two browser tabs. Perplexity's online models can pull fresh results without leaving the chat.
Custom AI Agents
Agents work like task-specific assistants: you set a name, a description, instructions, a welcome message, and optional PDF knowledge, then reuse the setup for every similar job. The library ships with presets for cover letters, SEO scoring, YouTube outlines, and travel plans. Grounding an agent in your own documents is the part that saves the most time, because you stop re-explaining context.
AI Image Generator Across Several Engines
Prompts turn into images through multiple engines, with controls for aspect ratio, style, and model choice. Generated files land in an image gallery, where you can download them or reuse a prompt with a different engine. For social work, the practical win is speed: a rough visual for a post takes seconds, not a designer's afternoon.
AI Video Generator and Presentation Builder
Video generation is available on the upper plans, with Pro capped at roughly five videos a month. The presentation tool turns an outline into slides, which is handy for a client update you'd otherwise build by hand. Output quality varies by model, and long or complex scenes still need editing elsewhere.
PDF Analysis and Image Reading
Upload a PDF or a screenshot and ask questions about it. The chat answers with references to the section it used, which makes fact-checking faster than scrolling a 60-page report. Text-based PDFs work far better than scans, and the docs suggest models such as GPT-4o, Claude Sonnet, and Gemini for this job.
Web Search and Deep Research
Switching on web search adds live sources to an answer, and Deep Research runs a longer investigation that returns a structured brief with links. That combination covers the gap between a quick lookup and a proper research session. Citations are visible, so you can check a claim before you publish it.
Prompt Library, Folders, and Custom Instructions
The prompt library holds hundreds of ready-made prompts, and custom instructions keep your tone consistent across chats. Folders sort conversations by client, project, or content type, with drag and drop and color labels. On a busy account, that organization matters more than any single model.
Pros and cons
Pros
- One bill for a wide stack: text, images, video, search, and document work all draw from a single AI credits balance.
- Model choice in place: you can answer the same question with GPT-4o, Claude, and Gemini without paying three subscriptions.
- Agents grounded in your files: upload a PDF as knowledge and the agent answers from your material, not just general training data.
- Annual billing cuts 20%: Starter drops from €19 to €15.83 a month, and the same discount applies across the four tiers.
- Organization that survives volume: folders, a prompt library, and custom instructions keep months of work searchable.
Cons
- Credits are hard to forecast: premium models and image runs cost more, so a busy week can exhaust the monthly allowance before the renewal date.
- No public API: you can't wire the platform into your own product or automate calls from a script, which rules it out for developer workflows.
- Browser only: a native iOS and Android app is still in development, so mobile work depends on the mobile web version.
- Video is limited: the AI video generator is capped on lower tiers and still trails dedicated video tools on control and length.
- The rebrand causes confusion: support pages, old reviews, and the bluegpt.app domain all mix the two names, which slows down troubleshooting.
Frequently asked questions
Both names point to the same product. The company renamed BlueGPT to VelvetYuzu in 2026, and bluegpt.app now redirects to velvetyuzu.ai. Your account, credits, and chat history carry over under the new name.
Related content
Explore related tools, skills, and articles for BlueGPT.
BlueGPT Alternatives
Odyssey 3
Odyssey · OtherOdyssey-3 is Odyssey's foundation world model. It simulates physics in real time, turns a prompt into an interactive environment, and adapts to robots, cars and games.
JarvisCore
Prescott Data · Coding · Productivity · OtherJarvisCore is an open source Python runtime where AI agents work as equal peers on an agent mesh, claim tasks from a shared ledger, and reach external services through a zero-trust broker.
Clef
Cloudflare · Business · OtherClef is Cloudflare's open-source decision model for Workers AI. It returns calibrated probabilities for typed questions, not generated text, so agents can classify, route, and act fast.