AIDA — “AI: Do Anything!” is my experiment in democratizing GenAI for non-technical people: simple, in-browser actions with BYOM (bring your own model: OpenAI, Anthropic Claude, Google Gemini). Keys stay local; no vendor lock-in.
If you’re new to AIDA, start here: https://blog.enree.co/2025/09/aida-v0-6-2-images-context-menus-bias-detection-and-a-smoother-everyday-ux
TL;DR
- New in 0.6.3:
🔔 Notification badge + Messages & Updates modal • Upload & Analyze Image (drag-&-drop, 10MB validation, optional question) • Multi-configuration Settings (save multiple provider setups and activate one) • Knowledge Base (embeddings) — preview: upload files, generate vectors, browse and manage (integration across actions coming next).
What’s new in AIDA v0.6.3
1) Notifications that matter
- A new bell icon shows an unread badge with counts.
- Click to open the Messages & Updates modal for release notes, tips, and important announcements.
- Read/unread state persists across sessions.
2) Upload & Analyze Image (from the popup)
- You’re no longer limited to images on a page: drag-and-drop or choose a file directly in the popup.
- Optional question field for targeted vision queries.
- File validation (common formats, up to 10MB), clear success/error toasts, and a spinner while enqueueing the job.
- Results appear in Results, with local file path context.
3) Multi-configuration Settings (per provider)
- You can now store multiple configurations (one per provider) and activate one at a time.
- Cleaner model pickers and better status feedback.
Heads-up after updating: visit Settings and re-enter/refresh your keys and model selections to align with the new configuration system.
4) Knowledge Base (Embeddings) — preview
- Upload text files (.txt, .md, .json, .csv, .html, .xml) to build embeddings with your chosen provider (OpenAI or Google for now).
- Add description and tags; process to create vector representations; browse stats (files, chunks, size) and filter by provider/model.
- Not wired into all actions yet — I’m exploring how to best expose KB-aware workflows (e.g., personalize “LinkedIn comment” suggestions with your style, documents, and rules).
How I use it (real story)
AIDA lives in my browser, so low-friction tasks stay low-friction. For instance, I often use Ask Anything to draft a LinkedIn comment while I’m on the post — no copy/paste into a separate tool, and the model has page context. That saved seconds add up. This will be increadibly useful by adding internal knowledge base and help the LLM to know your style.
How to try the new stuff
Upload & analyze an image
- Open AIDA → Other Actions → Upload & Analyze Image
- Drop a file (≤ 10MB) and optionally type a question
- Click Analyze Image → check Results
Right-click workflows
- Analyze image on page: Right-click → Analyze with AIDA
- Summarize selection: Highlight text → Right-click → Summarize with AIDA
- Sentiment & Bias: Highlight text → Right-click → Sentiment & Bias
Knowledge Base (preview)
- Settings → Knowledge Base tab → Select File
- Add description/tags → Process File for Embeddings
- Browse stats and entries; integration across actions is coming soon
Roadmap
- KB-aware actions: use your embeddings to personalize summaries, Q&A, social replies, etc.
- Logging & stats (local): understand which LLMs/models you invoke and how often (privacy-respecting).
- More on my mind but not yet clear!
Have suggestions or want to contribute ideas? I’m building this in the open—your feedback shapes what ships next.
Get started (BYOM)
Pick a provider in Settings (OpenAI / Anthropic / Google), paste your API key, and choose models. Keys stay on your device.
New to Gemini? There’s a free tier—great for testing.
👉 Landing page: https://organizer.solutions/briefly.html