The DocPilot Blog
Notes on AI documentation from the team building it.
Playbooks for engineering leaders, deep dives into code intelligence, and lessons from running the Onzira Labs platform in production.
vector-databases
The Complete Guide to Vector Databases for Documentation Search
Keyword search fails when users don't know the right words. Vector databases fix that — here's how to use them for documentation search that actually understands intent.
careers
The Rise of the AI Documentation Engineer Role
A new engineering role is emerging at the intersection of AI, developer experience, and technical writing — and hiring managers are scrambling to define it.
ci
How Modern Docs Enable Continuous Integration for Non-Code Assets
Documentation is code that ships. Here's how to wire diagrams, runbooks, and architecture specs into the same CI pipeline that catches broken imports.
cost
How We Cut Docs Generation Costs by 80% Using Local Embeddings
Cloud embedding APIs were burning our budget. Here's the engineering breakdown of how we cut costs 80% by going local—without sacrificing retrieval quality.
semantic-search
The Rise of Semantic Code Search: Trends for 2026
Semantic code search is moving from novelty to infrastructure. Here's what engineering teams should expect, build, and budget for in 2026.
websockets
Server-Sent Events vs WebSockets for Streaming AI Responses
Both protocols can stream tokens from an LLM, but they make very different trade-offs. Here is how SSE and WebSockets compare on the parts that decide on production.
postgres
Why We Chose Postgres Over Redis for Our AI Job Queue
Redis looked like the obvious pick for our AI job queue. Postgres ended up being the right one — and the reasons had nothing to do with speed.
github
Building a GitHub App: A Complete Developer Guide
A practical, code-first walkthrough that takes you from an empty repo to a deployed GitHub App handling webhooks and authenticated API calls.
postgres
Postgres + pgvector: Building Vector Search for Documentation
You don't need a separate vector database to make documentation searchable by meaning. Here's how to build semantic search on top of Postgres using pgvector—and what to watch out for once you scale.
search
Semantic Search vs Full-Text Search for Developer Docs
Full-text search finds your words. Semantic search finds your meaning. Most developer docs need both — here's how to decide which to invest in first.
rag
How to Use RAG to Answer Questions About Your Codebase
Most AI assistants hallucinate when asked about your private repo. RAG fixes that by grounding answers in your actual files — and it is simpler to set up than you think.