Free AI Toolkit (2/3): Images, Video, Voice and No-Code Apps With No Paid Subscription
Faisal Adnan · 05 Nov 2026
This is the second post in a three-part list of free AI tools. The first covered running models and agents on your own computer. This one covers the creative and no-code side: apps you can assemble visually, images, video and speech. As before, tool names are plain text and licences are noted.
Key takeaways
- Dify, Langflow and n8n let you build chatbots and AI workflows visually, without much code.
- Whisper can transcribe and translate speech, including Urdu, and faster-whisper runs it faster with less memory.
- Model weights can carry separate terms from the code. Read the model card before using a model in a product or a paid project.
No-code AI apps
- Dify (Dify Open Source License, Apache plus conditions): build chatbots and AI workflows visually.
- Langflow (MIT): a drag-and-drop builder for agents and RAG, with built-in API and MCP servers.
- n8n (fair-code, not OSI-approved): automate work across apps, APIs and AI, with a visual canvas and custom code when you need it.
Image generation
- ComfyUI (GPL-3.0): a node-based visual editor for diffusion models, very flexible.
- Fooocus (GPL-3.0): a one-click, Midjourney-style generator that works offline.
- InvokeAI (Apache-2.0): a polished creative studio for AI images.
Video generation
- Wan2.2 (Apache-2.0): text or image to video with open models.
- HunyuanVideo (check the licence): Tencent's open text-to-video model.
- LTX-2 (check the licence): fast video with synchronised audio.
Speech
- Whisper (MIT): transcribes and translates speech, including Urdu.
- faster-whisper (MIT): the same model running up to four times faster with less memory.
- Piper (GPL-3.0): lightweight offline text-to-speech voices.
The warning every builder should read
Model weights can carry separate terms from the code. A project's code can be MIT-licensed while the model it downloads has a different licence that limits commercial use. Always read the model card before using a model in a product or a paid project.
Licences in plain words: MIT and Apache-2.0 are permissive. GPL-3.0 requires you to share changes if you distribute the software. "Fair-code" and custom licences add conditions, so read them before you build a business on top.
A first project for each
- Whisper: transcribe a lecture recording in Urdu and clean the transcript.
- n8n or Dify: make a bot that answers questions from your department's handbook.
- ComfyUI or Fooocus: generate illustrations for a school project, and credit the tool in your report.
These tools need a decent computer for images and video. If yours is slow, start with speech and no-code tools and leave video for later. When you create images or video, remember the lessons from Is That Real? Increasingly, No: label what is AI-made and never use someone's face or voice without consent. The last post in the toolkit series covers testing, security and learning from the source.
Keep reading
- Free AI Toolkit (1/3): Run AI on Your Own Laptop and Build With It
- Free AI Toolkit (3/3): Test It, Secure It, Learn It From the Source
- Type a World, Walk Into It: Genie 3 and What AI Means for Game Development
See the complete AI Era series
Written by Faisal Adnan
Business Intelligence and AI mentor. I teach Data Analytics, Power BI, Microsoft Fabric, SQL, Python and applied AI, and write about what is changing and what it means for your career.