Writing

Ideas, updates,
and hard-won lessons.

Engineering posts from the team building n00dles, plus every version we've ever shipped.

Engineering
Parallel agents with one line of code
Most LLM pipelines are embarrassingly sequential. Here's how n00dles' parallel() primitive works under the hood.
Ryo Tanaka · May 316 min →
Production
State management in production LLM pipelines
Checkpointing, distributed locks, crash recovery, and why your in-memory state store will eventually betray you.
Leo Marchetti · May 2211 min →
Opinion
The hidden cost of LangChain
It's not just the boilerplate. The real cost is the abstraction debt you accumulate every time you paper over a bad API.
Zara Okonkwo · May 87 min →
Tutorial
Building a deep research pipeline in 50 lines
Web scraping, parallel analysis, synthesis, and structured output — all wired with n00dles in one afternoon.
Priya Sharma · April 2914 min →
Deep dive
How we think about retries in multi-agent systems
Not all failures are equal. Transient errors, rate limits, semantic failures, and timeouts each need their own strategy.
Leo Marchetti · April 148 min →
Release
Announcing n00dles v0.1.0 — public beta
After six months of internal development and two production deployments, we're open-sourcing n00dles.
Zara Okonkwo · April 15 min →
v0.1.0
Jun 2026
Public beta — open source release
First public release. Everything is stable enough for production use. Expect breaking changes before v1.0.
  • NEW@agent decorator with full type validation
  • NEWpipeline(), parallel(), branch() primitives
  • NEWSupport for Anthropic, OpenAI, Mistral, Gemini, Ollama
  • NEWPersistent state store (Redis + SQLite adapters)
  • NEWStructured trace logging with OpenTelemetry export
  • NEWnoodles deploy CLI command
v0.0.9
May 2026
Parallel execution + Pydantic validation
Rewrote the executor to use a proper async task graph instead of sequential coroutines.
  • NEWparallel() with automatic fan-out and result merging
  • NEWPydantic v2 models as agent output types
  • IMP40% reduction in per-agent overhead
  • FIXState deserialization edge case with nested dicts
  • FIXTimeout not propagated correctly in nested pipelines
v0.0.8
Apr 2026
Retry engine + circuit breaker
  • NEWExponential backoff with jitter per agent
  • NEWCircuit breaker — opens after N consecutive failures
  • NEWPer-provider rate limit detection and backoff
  • BRKon_error callback signature changed — see migration guide
v0.0.7
Mar 2026
Initial internal alpha
First version used in production. Sequential pipelines only, no parallel support. But it worked.
  • NEW@agent decorator (sequential only)
  • NEWBasic >> pipeline chaining
  • NEWAnthropic and OpenAI providers