Protobuf has LSP support

📝 Discussion Summary (Click to expand)

Summary of the three most prevalent themes

Protobuf’s relevance is being questioned in the LLM era
“Buf’s offering of protobuf registries and codegen SDKs for microservices seems less necessary in the LLM era” — echelon
Strict contracts are seen as essential when LLMs generate code
“When you have probabilistic models generating code, strict contracts become more critical, not less” — sudorandom
Practical concerns over ceremony, performance, and tooling persist
“If you patch clients to google services in Python to use json instead of grpc they get faster and more reliable” — ltbarcly3

🚀 Project Ideas

Eliminate manual protoc steps by letting developers describe schemas in simple markdown.
Auto‑generate type‑safe client libraries and an LSP that works directly in editors.

Key	Value
Target Audience	Small‑to‑mid‑size dev teams tired of protobuf tooling overhead
Core Feature	Natural‑language schema authoring → on‑the‑fly LSP & client code generation
Tech Stack	Node.js backend, Tree‑sitter parser, LLM‑driven codegen, Vue/React LSP client
Difficulty	Medium
Monetization	Revenue-ready: {SaaS subscription: $15/mo per seat}

HN users repeatedly lamented “ceremony” and “build steps” around protobuf; this removes both.
Quote from sudorandom: “the ceremony around generating code is also performed very well by LLMs.”
Potential for quick adoption by teams already using LLMs for code assistance.

Provide a declarative schema format (based on JSON Schema) that validates data contracts across services.
Auto‑generate OpenAPI/Protobuf equivalents to keep compatibility without manual versioning hassles.

Key	Value
Target Audience	Backend engineers managing multiple micro‑services with evolving schemas
Core Feature	Schema authoring in markdown → schema validation + automatic contract docs
Tech Stack	Python (FastAPI), Pydantic, JSON Schema, Docker, GitHub Actions CI
Difficulty	Low
Monetization	Hobby

Commenters emphasized the need for “single source of truth” without protobuf’s build steps.
Quote from cyberax: “Anybody who thinks you can just chuck unstructured data into LLM … is an idiot.” This tool offers structure without heavy protobuf.
Appeals to teams wanting deterministic layers while still using LLMs for generation.

Serverless API that accepts protobuf payloads, converts them to optimized JSON, and caches results.
Handles schema evolution gracefully, removing the need for separate JSON endpoints.

Key	Value
Target Audience	Companies with existing protobuf/gRPC pipelines seeking faster, simpler APIs
Core Feature	Real‑time protobuf → JSON conversion with intelligent caching and schema versioning
Tech Stack	Go (gRPC‑go), Redis cache, Cloud Functions (AWS Lambda), OpenAPI spec generation
Difficulty	High
Monetization	Revenue-ready: {Pay‑per‑request tiered pricing}

Several users pointed out “protobuf is slower than JSON” and suggested using JSON when possible.
Quote from blanched: “protobuf is faster at both encoding and decoding than json… but JSON’s advantages are not performance.” This service bridges that gap.
Offers a pragmatic path for teams reluctant to abandon protobuf but wanting leaner front‑end consumption.