Ho provato 12 Tool AI GRATIS Nuovi (Distruggono quelli a Pagamento)

Overall premise / evaluation

• The creator claims they tested many free AI tools over months and argues that 12 free options can outperform paid alternatives, potentially saving hundreds of euros.
• The video is organized into categories: brains (LLMs/search), creatives (images/design/presentations), builders (apps/automation), and surprises (music + advanced Google media tools).
• Repeated theme: free plans are surprisingly capable, but limitations often relate to privacy (VPN), upload/file size restrictions, or free-tier credit caps.

---

Block 1 — “Brains” (LLMs + research)

1) DeepSeek (“Deep PSC” in subtitles)

• Free + unlimited chat via web/app: no paywalls, no message limits (as stated by the creator).
• Uses a unified model (V3.2) that automatically switches between:
  • fast chat
  • deep reasoning
• Performance highlights:
  • Strong reasoning performance (as emphasized).
  • Mentions V3 Special allegedly winning gold in the 2025 International Mathematics/CS Olympiad.
  • Rumored V4: claims of ~1000B parameters, up to ~1M token context, and multimodal capabilities (text/images/video in one model).
• How to use now:
  • Access through VPN (privacy concern; servers in China mentioned), then deepsic.com.
  • Supports file/document upload for complex questions.
• Developer bonus:
  • 5M free tokens for API startup; pricing described as much cheaper than OpenAI.
• Best fit: complex reasoning and logic problems.
• Main downside: privacy + need for VPN.

2) Qwen (“Quen 3.6”)

• Presented as an all-in-one free model from Alibaba, with:
  • A single interface for text + images + videos + documents + coding + advanced search + artifacts
  • No obvious limits/paywalls (as described for the free tier)
• Practical demo workflow:
  • Upload a PDF → it extracts/breaks down key points.
  • In the same conversation → generate an image cover, then landing page code.
• Language support emphasized: 200+ languages/dialects including Italian.
• Mentions:
  • Omni capabilities (audio/video durations).
  • OmniPlus scoring high in “art scores” (metric not clearly specified).
• Best framing: do “everything” without paying CGPT Plus-like subscriptions.

3) Perplexity

• Positioned as a chat + search hybrid rather than classic LLM chat.
• Strength: answers include inline sources/citations directly under statements.
• Free plan (as claimed):
  • Unlimited/basic searches
  • 5 Pro searches/day for deeper multi-step reasoning
• “How to use best” tips:
  • Treat it like an analyst, not Google: ask specific, complex questions.
  • For video prep, request complete overviews with updated data.
  • In free tier, you can paste URLs/text for analysis.
• Limitations vs Pro:
  • Cannot upload large files.
  • No access to some premium models (as referenced: latest Claude/“Cloud”/GPT models).
• Outlook: the creator mentions computer mode potentially becoming powerful for agentic use.

---

Block 2 — “Creatives” (images/design/presentations)

4) Leonardo AI

• Free tier described as 150 tokens/day, resetting every 24 hours.
• Token economics:
  • Example: one standard image uses about 5–8 tokens
  • Roughly 20–30 images/day
• Key free-tier trick:
  • “Relaxed Generation mode” (claimed) allows indefinite generation (slower queue).
• Creative/product features:
  • Multiple models (e.g., Phoenix, Flux, others).
  • Canvas editor for post-generation edits:
    • inpainting/outpainting, background removal, upscaling, etc.
  • Real-time canvas:
    • hand-drawn sketch → transforms into a professional image while you watch
• Commercial rights (as stated):
  • Free use includes a non-exclusive commercial license
  • Private images require a paid plan
• Differentiation: framed as the practical + affordable option.

5) Ideogram (Idogram)

• Niche strength: highly readable text inside generated images (logos/posters/social graphics).
• Creator claim: ~85% accuracy for correct, legible typography (higher than typical generators).
• Free plan:
  • 10 slow credits/week (reset weekly)
  • Each prompt generates 4 images → up to ~40 images/week
• Practical use:
  • Generate a logo with correct text
  • Prompt with style cues (e.g., “typographer”, “poster”)
• Suggested replacement: can substitute for Canva for many use cases.
• Pricing hint: Basic plan ~€8/month (as stated).

6) Gamma App

• Claimed traction: 50M+ users.
• Core feature: create presentations/documents/landing pages from a prompt in <60 seconds.
• Free onboarding credit:
  • 400 free credits per new user
  • Estimate: ~40 credits per full presentation + ~10 credits per modification
• Strengths from demos:
  • Fast generation of structure, slides, visuals
  • Strong customization and quick updates
• Exports:
  • PDF, PowerPoint, PNG, Google Slides, etc. (not locked in)
• Automation:
  • Mentions API generation and 8,000+ integrations (Zapier/Make-style)
• “Import” feature:
  • Upload existing PDF/Word/PowerPoint/URL → Gamma restructures and you can modify, then export
• Framed as replacing Canva Pro for these workflows.

---

Block 3 — “Builders” (app + workflow creation)

7) Google Gemini Studio — “Build”

• Treated as “Gemini on steroids” for building apps.
• Build flow:
  • Gemini Studio → Build
  • Choose the model, add prompts, enable “supercharge” Google integrations/features
• Demo:
  • Prompt to generate an app that creates Instagram hooks in a slot-machine style
  • Emphasis: output is super functional
• Prompt-building help:
  • “Discovery/Remix App” gallery of working apps/workflows
  • Or ask Gemini to write a high-quality build prompt following best practices
• Positioning:
  • Replaces Bolt/Lovable/Replit-like tools for many simpler apps
  • Complex apps still need code/platform tools

8) Google Opal (automation competitor)

• Presented as simpler than n8n/Zapier.
• Access via opal.google; may require VPN depending on region.
• Key value:
  • Integrated with Google suite and free/simple
  • Build automation workflows using English instructions with a visual graph (nodes/triggers)
• AI agent mode:
  • Mentions a “mode function” (Feb 2026) using Gemini 3 Flash to plan/execute multi-step tasks
  • Asks the user only if info is missing
• Use cases described:
  • Internal operations workflows like onboarding, writing, brainstorming, data collection, data analysis
  • Works with Gmail/Docs/Sheets/Calendar
• Difference vs Gemini Studio Build:
  • Opal focuses on internal workflows/automation, not publishable web apps

9) NotebookLM

• Called “game-changing,” referencing earlier videos.
• Free limits (as stated):
  • Up to 100 free notebooks
  • Up to 50 sources per notebook
  • Up to 500 words per notebook
• Source types:
  • PDFs, Google Docs/Slides, YouTube videos, audio, etc.
• Key feature emphasis:
  • Audio/video overview / slideshow options (on-screen)
• Best-practice tip:
  • Configure output before generation using settings/custom instruction panel
  • Important because large context can dilute results
• Demo outcomes:
  • Generate a two-voice podcast from audio review settings
  • Support for Italian output, even when input info is in other languages
  • Turn many source materials into concise team-ready slide summaries
  • Mentions generating “videos” too (described as cute outputs)
• Positioning: replaces many apps by consolidating analysis + summarization + presentation generation.

---

Block 4 — “Surprises” (music + advanced Google media tools)

10) Suno AI

• Presented as uniquely able to generate complete songs (vocals + instruments + arrangement/mix) from a text prompt.
• Free trial:
  • 50 credits/day, resets at midnight
  • Each generation produces two variants, costing ~10 credits
  • Creator estimate: about 10 songs/day
• Two creation modes:
  • Simple mode: describe the song in natural language + lyrics
  • Custom mode: specify genre/mood/tempo and more details
• Workflow:
  • Generate <2 minutes, listen to both variants, then use continue to extend
• Important monetization note:
  • Free trial songs are personal use only; not for monetizing on Spotify/YouTube
  • Pro needed for commercial use

11) “Nano Banana Pro” (Google’s image generation tool: “Gemini + banana symbol”)

• Claimed to be state-of-the-art for image generation and easier than many competitors.
• Usage paths:
  • Inside Gemini (free) via the banana tool symbol
  • Also referenced via Google Studio (some functions may be paid)
• Recommendation:
  • Activate the “reasoning model” for better understanding
• Key advantage:
  • Understands simple natural-language instructions (compared to “chatting at a bar”), reducing the need for complex prompts
• Editing/iteration behavior:
  • Less likely to “lose context” across successive edits
• Advanced capability:
  • Combine a personal photo with a product context (for e-commerce scaling without reshoots)
• Prompt optimization fallback:
  • If you can’t craft complex prompts, ask Gemini to optimize into a 10/10 prompt
• Framed as replacing many other image generators when you want fast + high quality at free tiers.

12) Gemini Canvas

• Not just an editor—described as an AI feature that:
  • Transforms documents into presentations with AI-generated slides and visuals
  • Creates infographics
  • Converts material into a landing page (and can generate downloadable outputs like PDF)
• Suggested best practice:
  • Enable reasoning model, then use Canvas with prompts
• Claim: this capability remains available in the free model.

---

Key speakers / sources

• Main speaker/source: The video’s single creator/narrator (“M.” at the end of subtitles).
• Referenced companies/models/sources:
  • DeepSeek, Alibaba Qwen, Perplexity
  • Leonardo AI, Ideogram, Gamma
  • Google Gemini Studio (Build), Google Opal
  • NotebookLM, Suno
  • Google “Nano Banana” image tool”
  • Gemini Canvas