Vercel open-sources durable-harness, letting coding agents keep running after a crash
Vercel has open-sourced durable-harness, an npm package that can wrap coding agent frameworks such as Pi, OpenCode, fx, Claude Code, and Codex, writing sessions to a persistent log so that after a function crashes, a queue hands that turn to a new function to continue execution. Completed model and tool calls are read directly from the log and are not billed twice; tools can be marked as safe to replay or never to replay. The package is at version 0.0.x under the Apache-2.0 license, and can run on Vercel Functions plus Queues, on Node.js 22+ with Postgres, or on local files; Claude Code and Codex must run in Vercel Sandbox, and no performance benchmarks have been published.
Why read this
AI score 69
The original article clearly explains the crash-recovery logging mechanism, the replayable tool tiers, and the three runtime environments, making it easy to judge whether coding agents can be moved to the cloud.
Topics
Body
Add as a preferredsource on Google
Vercel has released durable-harness, an open-source package that keeps coding agent sessions alive through crashes and timeouts. Vercel durable-harness wraps existing harnesses, including Pi, OpenCode, Claude Code and Codex, without changing them. Each session lives in a durable log. A dead function no longer means lost work. For teams moving agents off laptops and into the cloud, that closes a real gap.
TL;DR
Size: npm package, version 0.0.x, Apache-2.0. Adapters cover 5 harnesses: Pi, OpenCode, fx, Claude Code, Codex.
Runs on: Vercel Functions plus Vercel Queues, Node.js 22+ with Postgres, or laptop files. Claude Code and Codex need a Vercel Sandbox.
Performance: No latency or throughput benchmarks published. All recovery claims are self-reported by Vercel.
Best: Finished model and tool calls are read back from the log, never re-run or billed twice.
Worst: Experimental 0.0.x API. Claude Code and Codex run only in a Vercel Sandbox, even self-hosted.
What is durable-harness?
durable-harness is a set of adapters from Vercel that store a harness session in a Workflow SDK World. A World bundles storage, a queue and a stream behind one interface. The package uses the same machinery that powers Vercel’s Workflow SDK.
Why do coding agents lose work in the cloud?
Harnesses like Pi and Claude Code were designed as long-lived local processes. They keep the session in memory. Locally, the workspace and history sit on the same disk, so resuming takes 1 prompt. In the cloud, a VM with persistent storage does not track which calls completed. Vercel team notes a failed turn or network error then means lost work.
How does durable-harness work?
It writes each step of a turn to a durable log. Before a model or tool call runs, it records an intent. When the call finishes, it records the outcome. Each turn is backed by a message on Vercel Queues. If the function dies, the queue hands the turn to a new function. That function reads the log and continues.
After a crash, the log shows what finished, what was running and what never started:
Completed calls are read from the log, so finished model tokens are not billed again.
replay: 'safe' tools, such as lookups, can run again.
replay: 'never' tools, such as refunds, never run twice. The model is told the call may have happened.
Harness session files are mirrored into the log as well. That covers Claude Code’s transcript, Codex’s rollout and OpenCode’s event store.
Where do files and shell commands run?
Model-chosen commands never run on the function’s own disk. The tools docs list 3 workspace options:
virtual() keeps files in the session log, with a bash-compatible shell and no network. It holds about 500 files.
virtual({ durable: false }) keeps scratch files in memory and clears them every turn.
sandbox() gives each session a Vercel Sandbox, with /workspace on a Drive.
Typed JSON documents, such as a to-do list, also live in the log. Each holds at most 64 KB.
Where can you run durable-harness?
The agent code stays the same across Worlds:
Vercel: the Vercel World, with every session traced in Agent Runs.
Local: files on your machine, with no configuration.
Self-hosted: the Postgres World, version 5.0.4 or later, on Node.js 22+. Vercel’s blog also lists MongoDB, Redis and Turso.
2 environment variables, WORKFLOW_TARGET_WORLD and WORKFLOW_POSTGRES_URL, move an agent to Postgres.
How does durable-harness compare?
What it isDurability adapters for existing coding harnessesDurable execution wrapper for OpenAI Agents SDKTypeScript multi-agent frameworkStateful agents on Durable Objects
Wraps Claude Code / CodexYes, experimental, Sandbox-onlyNot disclosedNot disclosedNot disclosed
Durability backendWorkflow SDK World (Vercel, local, Postgres)Temporal server or Temporal CloudInngest orchestration engineDurable Objects, Cloudflare Workflows
LanguageTypeScriptPythonTypeScriptTypeScript
LicenseApache-2.0MIT (server)Apache-2.0MIT
Status0.0.x, experimentalGA since Mar 23, 2026Not disclosedNot disclosed
Key Takeaways
durable-harness logs every model and tool call, then resumes crashed turns on a new function.
Finished calls are never repeated, so tokens are not billed twice.
Pi and OpenCode run in your function; Claude Code and Codex need a Vercel Sandbox.
Self-hosting works with 2 env vars and a Postgres World.
It is 0.0.x and experimental, with no published benchmarks.
Michal Sutter+ postsBio ⮌
Michal Sutter is a data science professional with a Master of Science in Data Science from the University of Padova. With a solid foundation in statistical analysis, machine learning, and data engineering, Michal excels at transforming complex datasets into actionable insights.
Michal SutterExa 3.0 Search API Tops Perplexity, Brave and Parallel on 8 Benchmarks
Michal SutterMeet Drex Max: Nace AI’s Multimodal Decision Model for Computer-Use Agents with a Score 67.59
RELATED ARTICLESMORE FROM AUTHOR
Google Cloud Ships llm-d 1.0 With Cache-Aware Routing for TPU and GPU Fleets
DeepSeek Open-Sources DeepInfer, the Inference Framework Behind DeepSeek-V4.1-Flash
Exa 3.0 Search API Tops Perplexity, Brave and Parallel on 8 Benchmarks