simonw

simonw / research-llm-apis

Public

Research into the HTTP APIs from various LLM providers.

18
1
100% credibility
Found Apr 05, 2026 at 18 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Shell
AI Summary

A set of example scripts that use web requests to capture and save raw responses from major AI providers, demonstrating features like chat, vision, tools, reasoning, and search.

How It Works

1
🔍 Discover AI Response Examples

You find a helpful collection of real-world examples showing exactly how popular AI chat services respond to questions, images, and tasks.

2
📥 Get the Example Files

Download the simple ready-to-run files that let you see these AI responses for yourself.

3
🔑 Link Your AI Accounts

Connect your accounts from AI services like ChatGPT makers, Claude, Gemini, or Mistral by adding your private access codes.

4
🚀 Run Your First Test

Pick a service and launch a test—it asks the AI questions about colors, weather, images, and more, feeling like watching magic happen.

5
Choose Next Adventure
💬
Basic Chat

See quick text replies to everyday questions.

👁️
Image Insights

Upload a picture and get a description.

🛠️
Smart Helpers

Watch the AI call pretend tools like weather checkers.

🧠
Deep Thinking

Observe step-by-step reasoning on tricky puzzles.

6
📁 Review Saved Replies

Open the new folder filled with raw chat logs, streams, and results to study side-by-side.

✅ Master AI Behaviors

You now have a personal reference library of AI responses, perfect for learning how they think and reply differently.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 18 to 18 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is research-llm-apis?

This GitHub research repo delivers shell scripts using curl to probe HTTP APIs from various LLM providers like OpenAI, Anthropic, Gemini, and Mistral. Set your API keys via env vars or the llm tool, run a script per provider, and it captures raw JSON/streaming responses for text completion, vision, tool calling, reasoning, web search, and beta features like extended thinking or MCP tools—dumping everything to a responses folder. It's raw github research data for understanding API quirks without SDKs.

Why is it gaining traction?

Stands out as a no-frills github research tool: pure curl hits real endpoints, revealing provider differences in streaming, multi-turn tools, and server-side search that docs gloss over. Developers grab it for quick prototypes or api surveys, skipping verbose client libraries. Pairs well with research github copilot workflows for dissecting responses.

Who should use this?

LLM integration engineers comparing providers for multi-model apps. Researchers building github research survey or paper prototypes needing sample payloads. Backend devs testing vision/tool flows before coding clients.

Verdict

Solid reference for LLM API research—clone, run, inspect—but low 1.0% credibility score and 18 stars signal early-stage maturity with thin docs. Worth forking if you're in github research group koch-style experiments; skip for production.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.