Skip to content

Nori Compare

Nori vs Goose: when does a team outgrow a local open-source agent?

Goose is a free, local-first open-source agent. Nori is a team cloud platform. Compares where agents run, Slack, schedules, models, shared context and cost.

Quick answer

Pick Nori if a team wants Jiro or any harness in isolated cloud sandboxes, started from Slack, schedules and webhooks, with shared memories and skillsets, on its own Claude or ChatGPT subscription and with no token charges. Pick Goose if one developer wants a free, Apache-2.0 agent on their own machine, with 85+ model providers and cron-scheduled recipes that run while that machine is on.

What is Goose?

Goose is an open-source, general-purpose AI agent with a desktop app and a CLI. Block created it, and it is now part of the Agentic AI Foundation at the Linux Foundation. The repository moved to aaif-goose/goose, and the old Block docs site points to the new one.

As of November 2026, Goose's own pages establish these points:

  • License and price. Goose is Apache-2.0 and free; no paid plan is listed (goose-docs.ai).
  • Models. Local models plus over 85 providers, including Anthropic, OpenAI, Gemini, OpenRouter and Amazon Bedrock. ACP agents such as Claude Code and Codex can act as providers.
  • Where it runs. Goose Desktop runs its own goose serve process on the same machine. You can also run goose serve on a remote VM and point the desktop app at it.
  • Schedules. goose schedule add --cron runs a recipe on a cron expression on the machine that hosts Goose.
  • Slack. A first-party Slack integration is not listed in the docs index.

What is Nori?

Nori is a cloud platform for coding agents, built for engineering teams. Work starts where the team already talks. Someone writes @nori the nightly export job fails on empty batches, find the cause and open a PR in a Slack thread; Nori starts an isolated cloud sandbox with the team's repositories, skillset and scoped credentials, and the pull request link comes back in the thread. The same sessions start from Discord, the web, the CLI, cron schedules, one-time reminders and webhooks, with one run history and approval paths.

Longer work lives in a project: a coordinator that remembers the goal, shared docs, and subworkers that each take a piece. Every session starts with the organization's skillsets (CLAUDE.md or AGENTS.md, skills, subagents), shared memories and integrations scoped to the org, a person or an agent. Nori runs any harness (Claude Code, Codex, Gemini CLI, Cursor, OpenCode, Droid, Amp or any ACP agent) on your own Claude or ChatGPT subscription, API keys or LLM gateway, and never charges for tokens.

Meet Jiro

Jiro is the AI engineering teammate that Nori ships. It runs on Claude Code or Codex with your own subscription or key and follows an engineering workflow: plan, write the failing test, implement, review. In its first days it reads your repositories, asks the team how you ship, and sets up the workspace. See jiro.bot.

Nori vs Goose, head to head

Nori vs Goose, as of November 2026. Goose facts come from its docs and repository, listed under Sources. "Not documented" means the pages we read do not establish it, not that Goose lacks it.
CategoryNoriGoose
HarnessJiro (Nori's own agent) or any harness: Claude Code, Codex, Gemini CLI, Cursor, OpenCode, Droid, Amp or any ACP agentGoose itself; ACP agents such as Claude Code and Codex can serve as providers
Where work runsAn isolated cloud sandbox per sessionYour machine, or a remote goose serve you operate
Ways inSlack, Discord, web, CLI, cron schedules, one-time reminders and webhooksDesktop app and CLI; first-party Slack not documented
SchedulesCron schedules, one-time reminders and webhooks with run history and approval pathsCron-scheduled recipes on the host running Goose
Model usageYour Claude or ChatGPT subscription, API keys or LLM gateway; Nori never charges for tokensLocal models and 85+ providers
Shared contextOrg skillsets, shared memories, scoped integrations, projects with shared docsRecipes and configuration on each machine; team-wide memory not documented
Pricing30-day free trial with no card; Developer $99/month (one person, three machines, +$50/month per extra machine); Team from $250/month (five machines, unlimited members); Enterprise customFree, Apache-2.0
Control planeHosted: a dedicated, single-tenant control plane and data store per organization. Enterprise: the full stack, control plane included, in your own AWS VPC or Azure VNet, or on-premNone to buy: you run Goose and own everything on that machine

Goose is software one person runs. Nori is a service a whole team uses. Most teams that compare them already have Goose working for someone and want the same help for everyone.

What changes when a team shares the agent

Running a local agent for a team means someone owns the machines, the keys and the schedules. A worked example shows where the work goes.

  1. Pick the job. Every weekday at 07:00, check the overnight CI failures on main and open one pull request per root cause.
  2. On Goose, write a recipe, add it with goose schedule add --cron "0 0 7 * * 1-5", and keep the host running with the repository, credentials and model keys on it. Results land where the recipe writes them.
  3. On Nori, create the automation below. It runs in a fresh sandbox with the team's skillset, posts to Slack, and appears in run history where anyone can rerun or approve it.
  4. Compare ownership. On Goose, the person with the host owns it. On Nori, the team does.
name:      weekday CI triage
schedule:  0 7 * * 1-5   (Europe/London)
agent:     Jiro
repos:     acme/platform
prompt:    Triage tests that failed on main overnight.
           Open one PR per root cause and post a summary.
deliver:   Slack #platform-ci

Enterprise deployment: where does the control plane run?

With Goose there is no vendor control plane: you run the software and decide where. That is full control, and it is also full operations work for the team that runs it.

Nori deploys the full stack, control plane included, into your own AWS VPC or Azure VNet, or on-prem, with your identity provider (SAML, SCIM, OIDC), your LLM gateway and your compute. On hosted plans, Nori gives each organization a dedicated, single-tenant control plane and data store, and every session runs in its own isolated sandbox.

When to choose Nori

  • The whole team asks for work. Slack and Discord threads start sessions, and anyone can follow or steer them.
  • Schedules should not depend on a laptop. Automations run in the cloud with history and approvals.
  • Context should be shared. Org skillsets and shared memories start every session the same way.
  • You want a teammate. Jiro plans, works test-first and reviews its own changes.
  • Security wants isolation and ownership. Each session gets its own sandbox, and Enterprise runs the full stack in your cloud.

Goose can also stay in the picture: engineers keep it locally while team work moves to Nori. See cloud coding agents compared.

When to choose Goose

  • One developer, one machine. A free local agent with a desktop app is the simplest setup.
  • You want open-source governance. Goose is Apache-2.0 under a Linux Foundation project.
  • You need local models. Goose runs local models directly, with no separate server.
  • You are happy to operate it. Remote goose serve and cron recipes work if someone owns the host.

How to move team work from Goose to Nori

  1. List your scheduled recipes with their cron lines, prompts, repositories and where they write results.
  2. Start the Nori trial. Thirty days, no card.
  3. Connect models with your Claude or ChatGPT subscription, API keys or gateway.
  4. Turn recipe instructions into skills in the org skillset, and standing rules into AGENTS.md.
  5. Recreate each schedule as an automation and point its output at a Slack channel.
  6. Run both for a week, compare results, then remove the cron entries from the Goose host.

Frequently asked questions

What is Goose?

Goose is a free, open-source AI agent with a desktop app and CLI. Block created it, and as of November 2026 it is part of the Agentic AI Foundation at the Linux Foundation, licensed Apache-2.0, with local models and over 85 model providers.

Can Goose run on a schedule?

Yes. The goose schedule add command runs a recipe on a cron expression, on the machine that hosts Goose, so that machine has to stay on. Nori runs cron schedules, one-time reminders and webhooks in the cloud with run history and approval paths.

Does Goose work in Slack?

A first-party Slack integration is not listed in the Goose docs we read in November 2026. Nori takes requests in Slack and Discord threads and posts results back to the thread.

Is Goose cheaper than Nori?

Goose itself is free, and you pay for your models and whatever machine runs it. Nori has a 30-day free trial, Developer at $99 per month and Team from $250 per month for five machines with unlimited members, and never charges for tokens.

Can I use Goose and Nori together?

Yes. Engineers can keep Goose on their own machines while scheduled and team work runs on Nori, which runs Claude Code, Codex, Gemini CLI or any ACP agent in isolated cloud sandboxes.

Sources

Goose pages were read on November 20, 2026. Open-source projects change quickly; check the source before you decide.