🧠 See AI First

The Opinionated AI Stack Guide — with receipts.

Curated directory of 66 AI developer tools across 13 categories, with evidence-first sourcing.

🌐 Live: seeaifirst.com

---

Why This Exists

There are hundreds of "awesome AI" lists. Most are link dumps with no opinion.

See AI First is different:

• Curated, not scraped. Every tool is manually evaluated against 5 criteria before inclusion. 100+ tools evaluated, 66 selected.
• Evidence-first. Each card includes sources, verified dates, and verification methods.
• Opinionated trade-offs. Every tool has whenToUse and whenNotToUse — we tell you when a tool is not the right choice.
• Structured for machines. Schema-frozen data with stable IDs, JSON-LD structured data — built to be referenced by AI agents, not just humans.

Features

• 🗂️ 13 categories across 5 layers: Foundation → Coordination → Capability → Application → Trends
• 🔍 Search across names, descriptions, and details (Ctrl+K)
• 🔗 Deep linking — every tool has a permanent URL via path-based routing
• ⚖️ Compare Mode — side-by-side tool comparisons with preset and custom selections
• 🎯 Tool Picker — interactive selection for building your AI stack
• 📊 Enriched metadata — pricing, deployment, difficulty, compatibility, use cases
• 🌗 Dark/Light theme
• 📱 Mobile responsive
• ⚡ Zero backend — static HTML + JSON on CDN, loads instantly

Usage

One effective way to use See AI First is to ask your AI agent to look things up for you.

Instead of browsing manually, you can ask your AI assistant (Claude, ChatGPT, Gemini, Grok, or any agent with web access) something like:

Locate a tool in the AI ecosystem:

• "Where does MCP fit in the AI stack? Check seeaifirst.com"
• "I found github.com/langchain-ai/langchain — where does this sit in the AI ecosystem? Look it up on seeaifirst.com"
• "Someone recommended Qdrant. What category is it in and what are the alternatives? Check seeaifirst.com"

Build your stack:

• "I'm building a RAG pipeline. What tools does seeaifirst.com recommend?"
• "Compare LangGraph vs CrewAI on seeaifirst.com"

Evaluate before adopting:

• "I'm considering pgvector for vector search. What does seeaifirst.com say about when to use it and when not to?"
• "What's the trade-off between LangSmith and Langfuse for observability? Check seeaifirst.com"

Validate social media recommendations:

• "I saw a post claiming FastCode is cheaper than Claude Code and Cursor. Is it on seeaifirst.com? Where does it fit in the coding agents category?"
• "Someone shared a new AI framework on Twitter. Check if seeaifirst.com covers it and what the trade-offs are"

The site is structured so web-enabled AI agents can read and summarize it more effectively — 66 tools mapped across 13 categories with opinionated trade-offs, so your agent gets context, not just links.

You can also browse directly at seeaifirst.com — use Search (Ctrl+K), Compare Mode, or Tool Picker to explore.

Quick Start

git clone https://github.com/BARONFANTHE/seeaifirst.git
cd seeaifirst
python -m http.server 8000
# Open http://localhost:8000

The site uses fetch() to load data, so a local server is required (opening index.html directly will fail due to CORS).

Alternative: npx serve .

Project Structure

seeaifirst/
├── index.html          # UI + CSS + JS (single file)
├── data.json           # 66 tools, 13 sections, 5 layers
├── og-image.png        # Social preview image
├── sitemap.xml         # SEO sitemap
├── robots.txt          # SEO robots
├── package.json        # npm config (for validator)
├── scripts/
│   └── validate.js     # Data validation (8 checks)
└── .github/
    └── ISSUE_TEMPLATE/
        └── suggest-tool.yml  # Community tool suggestions

Data

All tool data lives in data.json (English) and data.vi.json (Vietnamese). Each tool card includes:

• Basic info: name, description, sources
• Enrichment: pricing, deployment model, difficulty level
• Opinions: when to use, when not to use, compatible tools
• Verification: verified date, verification method

Schema is frozen at v1.0.0. See CONTRIBUTING.md for data format details.

Selection Criteria

Not every tool belongs here. Each suggestion is evaluated against 5 criteria:

Criteria	Minimum Threshold
GitHub Stars	Usually >5K ⭐ — exceptions allowed for innovative tools with strong rationale + evidence
Relevance	Must be part of the AI developer ecosystem
Maturity	Has documentation, active development, community
Uniqueness	Does not duplicate an existing card's functionality
Category Fit	Fits within existing 13 sections (new sections require strong justification)

100+ tools evaluated. 66 selected. See CONTRIBUTING.md for how to suggest a tool.

Contributing

We welcome contributions! The main ways to help:

• Suggest a tool → Open an issue
• Report a bug → Open an issue
• Improve data → Submit a PR (see CONTRIBUTING.md)

Tech Stack

• Frontend: Vanilla HTML + CSS + JS (single file, no framework)
• Data: JSON (schema v1.0.0 frozen)
• Hosting: Vercel (auto-deploy from main)
• SEO: JSON-LD structured data, canonical domain, sitemap + robots.txt

License

MIT

---

Vietnamese descriptions available on the live site.

Data verified as of February 2026. Found something outdated? Open an issue.