ChatGPT Gets New Features

Good morning, AI enthusiasts. OpenAI is giving free users a powerful new way to organize their AI workflows

PLUS: How to access MidJourney-level models for free

HOW TO AI

How to access MidJourney-level models for free

There’s now a website called Ghostbot Web that gives you completely free access to some of the best AI image models available today. With just a single prompt, you can generate high-quality images in less than a minute, and the results are surprisingly impressive. What makes this platform even better is that it provides free API keys so you can integrate these image models directly into your own apps, projects, or tools.

🧰 Who is this for

• Creators who want high-quality AI images without paying for free.

• Developers looking for free API access to integrate image generation into their apps.

• Designers testing multiple models for different styles and outputs.

• Anyone curious about AI art but not ready to pay subscription fees.

STEP 1: Access the tool

Go to Ghostbot.

Once inside, you’ll see two main options on the dashboard: “Text to Image” and “Image to Image.”

For this tutorial, click on “Text to Image.” This will open the image generation interface, where you’ll find a large text box labeled “Describe your vision.”

Here, type in the detailed description of the image you want to generate.

For example: “A massive dragon’s eye glowing with fiery orange and yellow light, its slit pupil sharp and menacing, dominating the scene. The scales around the eye are dark, rough, and ancient, lit by faint blue atmospheric light. In the foreground, a lone cloaked figure stands small and shadowed in the mist, facing the enormous eye, emphasizing the scale difference. The environment is dark, foggy, and cinematic, with a mystical, fantasy atmosphere. Ultra-realistic, highly detailed, dramatic lighting, epic fantasy concept art.”

STEP 2: Choose your image model

Below the prompt field, you’ll see a dropdown menu where you can pick from the top AI image models.

Options include models that normally require paid access elsewhere, such as Stable Diffusion XL or MidJourney-like models. Select the one that best fits your style preference.

You can also adjust settings such as aspect ratio. For example, choose 16:9 for wide cinematic images, 1:1 for Instagram-style posts, or 9:16 for vertical reels and stories.

STEP 3: Generate your image

Once your prompt and settings are ready, click the "Generate Image" button.

Ghostbot will process your request, and in less than a minute, your AI-generated image will appear.

You can download the image directly, refine your prompt for better results, or try different models to compare outputs.

The results are surprisingly detailed and often match or exceed paid platforms.

STEP 4: Use Free API Keys

If you want to take things further, go to the API Keys section in your account dashboard. Click “Generate New Key”, give it a name like “Image Generator Free Key,” and save it.

With this, you can plug the models directly into your own applications or tools. Whether you’re building a design app, a content creation pipeline, or an automation workflow, the API key lets you generate images programmatically.

TODAY IN AI

1. OpenAI Rolls Out ChatGPT Projects To Free Users

OpenAI is making its Projects feature available to everyone, including free ChatGPT users. Projects let you organize chats by topic, add files, and customize how the AI responds, essentially acting as a smarter folder system.

→ Free users can now upload up to 5 files per project

→ Plus subscribers get 25 files, Pro/Enterprise get 40

→ Projects include custom colors, icons, and memory controls

→ Rolling out on web & Android, iOS coming soon

Why is important: By moving a premium feature down to free, OpenAI is both expanding access and nudging casual users toward paid plans. It’s part of a broader strategy: release advanced features to paying users first, then democratize them later with limits, all while locking users deeper into the ChatGPT ecosystem.

2. Apple reportedly struck a deal with Google to power Siri’s AI search upgrade

Apple is preparing a major revamp of Siri with a new feature called World Knowledge Answers, an AI-powered web search tool set to launch in spring 2026. To make it happen, Apple has reportedly struck a deal with Google to integrate a custom Gemini model into Siri.

→ Siri’s “answer engine” will deliver summarized results from across the web

→ Apple is keeping its Foundation Models for on-device/private search

→ iOS 26.4 will be the first version to feature the new AI search

Why is important: Siri has been falling behind ChatGPT, Gemini, and Perplexity for years. This partnership signals Apple’s biggest leap in AI yet, turning Siri into a true rival while reducing dependence on Google’s search revenue long term.

ESSENTIAL BITES

• Perplexity made its Comet agentic browser free for students worldwide and partnered with PayPal and Venmo to offer Perplexity Pro and early Comet access to their customers.

• OpenAI employees can now reportedly sell roughly $10.3 billion in stock in increased secondary share sale.

• Nous Research released Hermes-4-14B, a compact hybrid reasoning LLM optimized for consumer hardware, improving math, code, STEM, logic, and creative writing performance on Hugging Face.

• MIT researchers created an AI model that predicts flu vaccine strains more accurately than WHO

• ElevenLabs launched an upgraded sound effects AI with looping, longer outputs, and better quality.

⚙️ Hot AI Tools:

• 🍌 Gemini 2.5 Flash Image Preview: Viral ‘Nano Banana’ AI editing tool

• 🕵️ Avoid.so: Bypass AI detection with 99.8% success.

• ⚙️ Grok Code Fast: xAI’s new SOTA coding model

• 📧 Ellie: Professional, bespoke emails, written in your style.

That's it! See you tomorrow

- Dr. Alvaro Cintas