Notes from the proxy.
No marketing fluff.
Posts on what we’re building, what we’ve learned running the proxy in production, and how to think about the LLM gateway category. Every post has a real human’s review on it. If a post sounds AI-written, it’s getting rewritten or pulled.
- Deep diveMay 26, 20262 min read
Vercel AI Elements: 20+ React components for AI apps explained
A walk-through of every AI Elements component, what each one solves, and where rolling your own still wins. Practical patterns, real composition.
Synapse Publication- how-to
- vercel-ai-sdk
- react
- ui
- How-toMay 22, 20262 min read
Vercel AI SDK chatbot tutorial: useChat, streaming, real patterns
A working production-grade chatbot built on Vercel AI SDK v6. Streaming with useChat, tool calls, persistence, and the patterns that hold up after the demo.
Synapse Publication- how-to
- vercel-ai-sdk
- tutorial
- react
- ComparisonMay 12, 20262 min read
OpenRouter alternatives in 2026: pricing, routing, self-host
A working comparison of LLM gateways that compete with OpenRouter. What each one is good at, the pricing catches nobody mentions, and how to actually pick.
Synapse Publication- comparison
- gateway
- openrouter
- tooling
- Field noteMay 10, 20262 min read
Why we built Synapse Garden
One API key, every model. The short version of how the proxy started, what it's for, and what we won't add to it.
Synapse Publication- company
- philosophy
- How-toMay 9, 20262 min read
How to switch from OpenAI to Claude without rewriting
A practical walkthrough of moving an OpenAI integration to Claude, Gemini, or any other model. SDK swap, prompt translation, tool-calling differences, streaming gotchas, and the parts that actually break.
Synapse Publication- migration
- openai
- anthropic
- how-to
- PillarMay 8, 20262 min read
Per-project API keys for LLMs — why one shared key fails
A single OpenAI key for the team is the default and the wrong default. What per-project keys actually buy you, what they cost to set up, and the failure modes a shared key creates.
Synapse Publication- security
- governance
- best-practices
- ComparisonMay 7, 20262 min read
OpenRouter vs Portkey vs Helicone vs LangSmith vs Synapse Garden
An LLM gateway comparison. What each tool actually does, where they overlap, and how to pick when your team can only commit to one.
Synapse Publication- comparison
- tooling
- observability
- gateway
- BenchmarkMay 6, 20262 min read
50ms of LLM proxy overhead — what it costs and buys
Every LLM gateway adds latency. The honest question is how much, when it shows up, and whether the tradeoffs are worth it. Real k6 measurements, not vibes.
Synapse Publication- performance
- benchmarks
- engineering