Build durable, observable, failure-resistant code with sandboxing, agent evals, and A/B tests built in.
Inngest is the platform for building durable workflows and agents that survive failures, without managing orchestration infrastructure. Write functions in your existing code and add steps. Inngest runs each step, saves its result, and retries failed work from the step that failed.
Quick start
Run your first durable workflow, Sandbox, eval, or live update.
inngest-kb.zunoywork.com
What is Inngest?
See how workflows, Sandboxes, and outcome scores fit together.
inngest-kb.zunoywork.com
Coding agents
Using a coding agent? Give it skills, MCP, and LLM-ready docs.
inngest-kb.zunoywork.com
SDK reference
TypeScript, Python, and Go SDKs, plus the REST API.
inngest.com
Durable execution, isolated Sandboxes, outcome scores, and live updates all share the same run, so you can follow the whole job in one trace.
Durable Execution
Coordinate steps, retries, waits, and events so work finishes after failures.
inngest-kb.zunoywork.com
Sandboxes
Run code in isolated microVMs as part of a durable run.
inngest.com
Agent Evals
Score outcomes and compare changes to steps and workflows.
inngest.com
Realtime
Stream progress from durable runs to your users.
inngest.com
Flow control
Control concurrency, throughput, and fairness across tenants.
inngest-kb.zunoywork.com
Observability
Inspect traces, replay runs, and track metrics in production.
inngest.com
For example, a support agent can call tools and run code in a Sandbox while it answers a request. Inngest retries failed steps and records the work in a trace. When feedback arrives later, a score links it to the run and experiment variant, so you can compare the outcomes of agent changes.
Explore use cases
See how teams use Inngest to run agents, background jobs, generated code, and experiments in production.
AI agents
Run agent loops that call tools, wait for people, and survive failures.
inngest.com
Workflows and background jobs
Run multi-step jobs that retry each failed step, not the whole job.
inngest-kb.zunoywork.com
Code execution
Run generated code and test repositories in isolated microVMs.
inngest.com
Evals and A/B tests
Score real outcomes and compare prompts, models, and workflows.
inngest.com
Start building
Pick your language to build your first durable workflow. Use Local Development to run and inspect it on your machine.
TypeScript
Works with Node.js, Bun, Deno, and edge runtimes.
inngest.com
Python
Sync and async functions with FastAPI, Flask, or Django.
inngest.com
Go
Typed events and durable steps with the standard library.
inngest.com
Next.js
inngest-kb.zunoywork.com
Express
inngest-kb.zunoywork.com
TanStack Start
inngest-kb.zunoywork.com
NestJS
inngest-kb.zunoywork.com
Astro
inngest-kb.zunoywork.com
H3
inngest-kb.zunoywork.com
FastAPI & Flask
inngest.com
Coding agent resources
AI development tools
Resources for using Inngest with AI coding agents, MCP, CLI workflows, and LLM-ready documentation.
inngest-kb.zunoywork.com
Agent Plugins and Skills
Official plugins and skills that teach coding agents how to write, test, and debug Inngest functions.
inngest-kb.zunoywork.com
Inngest MCP
Connect coding agents to Inngest Cloud or your local Dev Server to inspect functions, send events, and debug runs.
inngest-kb.zunoywork.com
LLM Docs
Full Inngest documentation in markdown format for LLM context windows. Also available as a table of contents.
inngest-kb.zunoywork.com
