AI Builder's Field Guide

A daily-curated wiki of the tools, models, and patterns that AI power-users and builders actually reach for.

Browse the live site → · RSS · Suggest a tool

---

What this is

A short, opinionated list of the AI tools, models, and frameworks that working builders actually use in practice — refreshed every day by an AI curator.

No marketing copy. No "revolutionary" adjectives. No five entries from the same vendor. Every link is verified live before it lands.

If you ship products on top of LLMs and want a fast scan of what's in use right now, this is for you.

What makes it different

• Every section tells you how to choose. The hard part isn't finding twelve coding agents — it's knowing which one fits your situation. Each category carries a short decision guide: the default pick, the two or three situations where something else wins, and the trade-off a vendor page won't mention.
• Curated by an agent, every day. A scheduled run re-verifies the oldest links, hunts for genuinely new tools with live web search, prunes what has gone stale, rewrites vague copy, and refreshes one decision guide. At most two additions per day — quality over volume.
• Use-case-first taxonomy. Categories are grouped into five sections — Build with AI, Models & Infrastructure, Build Agents, Generate Media, Measure & Monitor — so you can jump straight to what you actually need.
• Diff-able and forkable. The whole wiki is one JSON file (data/entries.json). Want to fork it for your own niche — clone, edit one file, deploy.
• Per-card permalinks + OG images. Every entry has a stable URL fragment and an auto-generated social card. Sharing a single tool to your team is one click.
• Plain static site. No JS framework, no tracking, no build server beyond the daily render. Loads in under a second.

What are you trying to do?

• I want AI in my editor → AI Coding Tools

• I want to run many coding agents at once → Multi-Agent Coding

• I want my AI-built UIs to stop looking generic → AI Design Skills

• I want a working app from a prompt → App Builders

• I need the best-available model behind an API → Hosted Frontier APIs

• I want a model I can run on my own hardware → Open-Weight Models

• I need to host an open model at scale → Inference & Hosting

• I'm building an agent product → Agent Frameworks & SDKs

• My agent needs to remember things → Agent Memory

• My agent needs to look things up → Retrieval & Search

• I need to give my agent tools → MCP Ecosystem

• I need a voice or audio output → Voice & Audio

• I need to generate or compose video → AI Video

• I need 3D assets — games, AR, products → AI 3D

• I need to know if my AI works → Evals, Benchmarks & Observability

Recently added

• Voiceflow — Voiceflow is an enterprise conversational AI platform for designing, testing, and deploying AI agents across chat and voice channels. (added 2026-11-03)

• Tencent Hy-SimulMT — Tencent open-sourced Hy-SimulMT, a model for real-time Chinese-English simultaneous interpretation with translated text and speech output, including real-time voice cloning. (added 2026-10-28)

• Hi3D V1.0 — Hi3D is an open-source model for generating high-resolution 3D geometry from images or text. (added 2026-10-27)

• PixVerse — PixVerse generates video from text or images, offering camera control and character consistency. (added 2026-10-26)

• Spira Maxima — Spira Maxima is an AI social media agent that generates, schedules, and publishes short-form videos and images from scripts using viral templates. (added 2026-10-18)

• Meta Muse Spark 1.3 — Meta Muse Spark 1.3 is a multimodal reasoning model from Meta, tuned for agentic workflows and coding, available through Meta Model API. (added 2026-10-12)

• AnySearch AI — AnySearch AI provides privacy-first search infrastructure for AI agents, offering structured access to high-quality information from diverse vertical domains beyond the public web. (added 2026-10-11)

• Google Gemini Agent — A unified AI agent from Google designed for workplace tasks, capable of operating across various Google and third-party applications. (added 2026-10-10)

Sections & categories

Build with AI

• AI Coding Tools (18) — Agents and IDEs that read, write, and run your code. Where you reach when you want AI inside your editor.
• Multi-Agent Coding (17) — Run, orchestrate, and embed fleets of coding agents — local, cross-device, or inside your product.
• AI Design Skills (5) — Reusable skills and prompts that teach coding agents about visual design.
• App Builders (11) — Generate a working full-stack app from a prompt. Prototype-to-production in minutes.

Models & Infrastructure

• Hosted Frontier APIs (7) — The closed-weight frontier models you call from your stack — Claude, GPT, Gemini.
• Open-Weight Models (9) — Downloadable model families. Host yourself or run on any inference platform.
• Inference & Hosting (12) — Where to run open-weight models in production — managed inference, fine-tuning, custom silicon.

Build Agents

• Agent Frameworks & SDKs (14) — The glue for building autonomous, tool-using agents.
• Agent Memory (11) — Persistent memory layers — what your agent knows across sessions, users, and turns.
• Retrieval & Search (7) — Vector stores, full-text search, and web search APIs. Stateless retrieval — different from memory.
• MCP Ecosystem (7) — Servers, clients, and registries for the Model Context Protocol — the standard for connecting agents to tools.

Generate Media

• Voice & Audio (5) — Speech synthesis, voice cloning, realtime agents, and music generation.
• AI Video (6) — Text-to-video models and programmatic video frameworks.
• AI 3D (2) — Text- or image-to-3D generators and 3D asset pipelines.

Measure & Monitor

• Evals, Benchmarks & Observability (16) — Measure what your agents and models actually do. Public leaderboards and production observability.

Suggest a tool

Open an issue with the URL and one sentence on why it belongs. The curator picks up suggestions on the next daily run.

For faster turnaround, open a PR editing data/entries.json directly. See CLAUDE.md for the entry schema.

Fork it

The whole site is a static-site generator over one JSON file. To run your own:

git clone https://github.com/felixkruger/ai-builder-wiki.git my-wiki
cd my-wiki
pip install -r requirements.txt
# Edit data/entries.json — keep the schema in CLAUDE.md
python scripts/render.py
# Enable GitHub Pages: Settings → Pages → Source: GitHub Actions
# Push to main — the deploy workflow ships it

For the daily curator, see docs/daily-curator-package.md.

How it's built

Piece	What it does
data/entries.json	Source of truth — every tool entry and category lives here
scripts/render.py	Renders index.html, feed.xml, OG cards, and this README's auto-sections
templates/	Jinja2 templates for the site, RSS, OG cards, README
skills/wiki-curator/SKILL.md	The daily curator's playbook
docs/daily-curator-package.md	Mon–Fri prompt for the scheduled Claude cowork session
.github/workflows/deploy.yml	Builds & ships to GitHub Pages on every push to main

License

Content under CC-BY-4.0. Code under MIT.