---
title: "Cursor Background Agents Alternatives & Competitors (2026): 7 Best Options"
canonical: https://jiro.bot/alternatives/cursor-background-agents
last_updated: 2026-11-06
---

# Cursor Background Agents Alternatives & Competitors (2026): 7 Best Options

Updated 2026-11-06

> **What are the best Cursor background agents alternatives?**
>
> For engineering teams between five and twenty people whose Claude or ChatGPT invoices are already paid, the strongest Cursor background agents alternative is Jiro; Claude Code on the web comes second, then OpenAI Codex cloud, Amp, GitHub Copilot's cloud agent, Tembo and Devin. Jiro is a background coding agent by Nori. Cursor now sells the feature as Cloud Agents from the $30-a-month Premium tier with Claude and GPT usage charged at API list price on top; Jiro instead runs Claude Code, Codex or Cursor's own agent in an isolated cloud environment on credentials you already hold, from $99 monthly, with nothing added to the token bill. It is not Amazon's Kiro or Atlassian's Jira.

## Cursor cloud agents and seven alternatives, side by side

|  | Agent it runs | Model supplier and billing | Where it runs | How work ships |
|---|---|---|---|---|
| Jiro | Claude Code, Codex, Cursor, OpenCode, pi-acp [1] | Your Claude/ChatGPT login or API key; Cursor key [2]; Your keys, zero token markup [3] | Isolated cloud environment with your repos and tools [4] | PRs, commits, comments for review [5] |
| Cursor cloud agents | Cursor's agent | Plan usage at the selected model's rate [6]; Prepaid credit at API list price; Teams add $0.25/M tokens [7] | Cursor cloud VMs or self-hosted machines [8] | PRs with screenshots and video [9] |
| Claude Code on the web | Claude Code | Claude models; Anthropic account required [10] | Anthropic-managed VMs [11] | You create the PR; auto-fix follows up [12] |
| OpenAI Codex cloud | Codex | OpenAI models; cloud needs a ChatGPT plan [13] | Codex-managed cloud VMs [14] | You open the PR [15] |
| Amp | Amp's agent | Own subscriptions and keys, no BYOK fee [16] | Amp or self-hosted orbs, or your own runners [17] | Pushes to trunk by default [18] |
| GitHub Copilot cloud agent | Claude and Codex as third-party agents [19] | Claude, Gemini, GPT, MAI via GitHub; no own keys [20] | Cloud sandbox, Actions (59 min) or remote env [21] | Draft PRs; cannot merge [22] |
| Tembo | Multiple harnesses; own keys or subscriptions [23] | Own keys or subscription; compute metered [24] | Tembo cloud sessions (Docker or VM) [25] | Opens PRs, including multi-repo PRs [26] |
| Devin | Devin's own agent; Desktop also hosts ACP agents [27] | Anthropic, OpenAI, Google, Cognition SWE; Adaptive routing [28] | Cognition cloud, or local CLI/Desktop [29] | Opens PRs and answers PR comments [30] |

Cursor and Jiro both let you pick the model; the difference is whose account the tokens bill to.

## Pricing and limits with units, checked 2026-11-06

|  | Cheapest plan with cloud agents | Price per unit | Included quota | Overage | Team pricing |
|---|---|---|---|---|---|
| Jiro | Developer | $99 per month, 1 user [31] | Up to 3 runtimes; model usage on your own plan or key | None from Nori; your model vendor bills tokens [3] | Team $250 per month, 5 shared runtimes |
| Cursor cloud agents | Premium | $30 per month [32] | Weekly usage pool; Premium 4x, Super 15x, Ultra 50x Plus [33] | Prepaid credit at API list price; Teams add $0.25/M tokens [7] | Free $0; Plus $8; Premium $30; Super $100; Ultra $200; Teams $30+/user [34] |
| Claude Code on the web | Claude Pro | In Pro $20/mo, Max $100, Ultra $200, Team seats [35] | Plan usage limits | Usage credits at API rates [36] | Team and Enterprise seats |
| OpenAI Codex cloud | ChatGPT Plus | In ChatGPT Plus $20/mo and up [37] | Shared with ChatGPT usage | Plan limits apply | Business $25 per user per month |
| Amp | Hobby (free) or Megawatt | Hobby free; Megawatt $20/mo; Gigawatt $200/mo [38] | $20 of agent usage and 750 orb hours on Megawatt | From $0.08/hour [39] | Teams: no extra charge; per-member tiers [40] |
| GitHub Copilot cloud agent | Copilot Pro | Pro $10; Pro+ $39; Max $100; Business $19/user [41] | AI Credits since 2026-06-01; 1 credit = $0.01 [42] | AI Credits at per-model rates | Business $19 per user per month |
| Tembo | Pro | Pro $60/mo ($90 usage); Max $200/mo ($400); VMs/sec [24] | $90 of usage on Pro | Cloud VMs metered per second | Pro up to 5 users; Max up to 10; 7-day trial [43] |
| Devin | Pro | Free; Pro $20/mo; Max $200/mo; Teams $40/seat, $80/mo min [44] | Plan quota; 10 concurrent on Free/Pro [45] | On-demand credits | Teams $40 per seat, $80 minimum |

Every figure appears in a fact cited on this page. Cursor's legacy Pro, Pro+ and Teams plans are priced separately and still honoured.

## What happened to Cursor background agents?

They were renamed. Cursor's help pages call the feature Cloud Agents, and the Cloud Agent docs confirm the earlier name was Background Agents [46] [47]. Each agent gets a cloud VM holding a complete development environment; alternatively, Self-Hosted Machines keep tool execution on hardware you control [8]. The environment is declared in `.cursor/environment.json` and built from a Dockerfile, from a snapshot you saved earlier, or by letting the agent configure itself; secrets belong in a Secrets tab [48].

The output remains the strongest part. A finished agent opens a merge-ready pull request in every repository it touched, with screenshots, video and logs attached [9]. Cursor also states that parallel agents are unlimited, publishing no concurrency ceiling [49].

So the product did not get worse. What changed is who it is sold to, and how.

## Why do teams look for a Cursor Cloud Agents alternative?

Price structure is the first reason. Subscriptions come only from X Subscription Corp, and both cloud agents and automations begin at Premium, $30 monthly [32]. Plus, at $8, offers Cursor's own Grok models at fixed settings and no Cloud Agents at all [50]. Premium's weekly usage pool covers Grok and Composer only [33]. Claude, GPT and Gemini sit outside that pool. From the very first request they are paid out of a prepaid balance at the provider's list price, and Teams and Enterprise accounts add a Cursor Token Rate of $0.25 for every million tokens on eligible requests [7].

For a team that already holds Claude Max or ChatGPT seats, that is paying for the same models twice. Customers on legacy Pro, Pro+ or Teams plans can stay, but a move onto the new plans cannot be reversed [51].

The second reason is control over where code runs and which agent runs it. Several alternatives below let you swap the harness, bring your own subscription, or both.

## How we ranked the alternatives

Five questions, in order. Whose account pays for model tokens, and is there a markup? Where does a task start: a chat thread, a GitHub issue, a schedule, a webhook? Does each session get its own environment with your toolchain installed? What comes back, and can a reviewer act on it without opening an IDE? And does team pricing scale by seat or by capacity? The ranking favours small groups, five to twenty engineers, whose Anthropic or OpenAI invoices already exist, because that group is most exposed to Cursor's separate billing for third-party models [7]. Solo developers on a Claude Pro plan may prefer the Anthropic and OpenAI options two and three places down.

### 1. Jiro: your subscription, your choice of agent, team pricing by runtime

Five harnesses are first-class in Jiro: Claude Code and Codex, Cursor Agent, then OpenCode and pi-acp [1]. Codex accepts OAuth or an OpenAI key; Claude Code accepts a setup token, a Claude sign-in or an Anthropic key; the Cursor harness wants a Cursor API key and nothing else [2]. Nori marks tokens up by nothing [3]. Developer ($99/mo) is for one person with up to 3 runtimes; Team ($250/mo) shares 5 runtimes across the group; the first 30 days are free and need no card [31].

Work begins in Slack threads, a web dashboard, Discord, Nori's Handroll terminal, cron schedules or inbound webhooks [52] [53]. Sessions each get an isolated cloud environment with the repository, dependencies and tools ready, prepared by an organisation-wide setup script [4] [54]. What comes back is a PR, or a commit, or a comment, for a human to review [5], and each session and action is logged [55]. At Miravoice, Nori produces more than 25% of merged pull requests [56].

Limits: GitHub only, via a GitHub App or PAT [57]. No screenshot or video artefacts on PRs.

### 2. Claude Code on the web: shortest hop if Claude was the point

Anthropic hosts each session on its own VM, 4 vCPUs, 16 GB memory and 30 GB storage by default, and keeps it running after you close the laptop [11]. The feature comes with Pro at $20 per month, then Max ($100) and Ultra ($200); Team and Enterprise seats include it as well [35], so a team already on Max pays nothing extra for the environment. Launch from claude.ai/code, from the Claude desktop and mobile apps, or via `claude --cloud` in a terminal, while `claude --teleport` brings a cloud session back locally [58]. The branch is pushed; opening the pull request is your job, and once it exists, auto-fix handles failing CI and reviewer comments [12]. It is Claude only and needs an Anthropic account; Bedrock and Vertex users are excluded [10]. Team and Enterprise Slack usage now goes through Claude Tag, which bills by usage [59] [60].

Best for: individuals and teams whose Cursor usage was mostly Claude.

### 3. OpenAI Codex cloud: bundled with ChatGPT from $20

ChatGPT Plus, $20 monthly, includes Codex cloud; Pro tiers cost $100, $200 or $500 monthly; Business is $25 per user monthly ($20 on annual billing), with limits shared with ChatGPT [37]. Each task gets an OpenAI-managed VM that keeps going while your machine sleeps [14]. Tasks begin inside ChatGPT on desktop, web or mobile; the Codex CLI can start them too [61]. Slack mentions now go to @ChatGPT [62]. Inspect the diff; commit or raise the PR once satisfied [15]. An API key covers Codex in the IDE and CLI only; cloud work needs a ChatGPT plan [13]. The older Codex Cloud (Legacy) still sits behind Code Review, Linear and GitHub, and OpenAI plans to retire it [63].

Best for: ChatGPT-first teams happy with OpenAI models only.

### 4. Amp: bring any subscription, mind the default Ship

Amp takes a Claude or ChatGPT subscription, other subscriptions, or raw API keys, and outside Enterprise adds no fee for tokens on keys you bring [16]; Claude plan support landed on 2026-10-10, built on Anthropic's Agent SDK [64]. Remote work happens in orbs: each thread gets one machine, either Amp-hosted or on infrastructure you operate [17]. Hobby costs nothing beyond pay-as-you-go orbs. Megawatt: $20 per month, including $20 of usage and 750 small-orb hours. Gigawatt: $200 per month [38]. Orb time is metered per minute, starting at $0.08 hourly [39].

Two cautions for Cursor users. Amp's Cursor and VS Code extensions went dark on 2026-03-05, leaving the CLI and apps [65]. And the default Ship rebases, runs the tests, then commits to trunk itself; flip Push to Branch to get a PR [18]. Surfaces include the web app, macOS, iOS and Android, the CLI, GitHub and Slack [66].

Best for: engineers who want BYO subscription and do not need an editor plugin.

### 5. GitHub Copilot cloud agent: cheapest entry, most trigger surfaces

Individuals pay $10 monthly for Pro, $39 for Pro+ or $100 for Max; organisations pay $19 per user on Business and $39 on Enterprise; the Free tier has no cloud agent [41]. Sessions run in GitHub's own cloud sandbox, in an Actions job limited to 59 minutes, or in a remote environment of yours [21]. All plans have metered usage in GitHub AI Credits since 2026-06-01, one credit being $0.01 [42]. Model choice is Auto or GitHub-supplied Claude, GPT, Gemini and MAI; bring-your-own-key is not documented for the cloud agent [20]. Starting points are the widest here: GitHub.com and GitHub Mobile, the IDEs, the CLI and REST API, and Azure Boards, Jira, Linear, Microsoft Teams and Slack [67]. It works on its own branch, opens a draft PR, and lacks any ability to approve or merge [22]. Claude and Codex can take GitHub assignments as third-party agents on paid plans (public preview) [19].

Best for: GitHub-centric teams who want the cheapest seat and a Teams or Jira trigger.

### 6. Tembo: harness buffet with per-second compute

Tembo offers Pi, OpenCode, Codex and Claude Code as per-task harnesses; billing goes to your Anthropic, OpenAI or xAI key, or to a ChatGPT, Claude or SuperGrok plan [23]. Sessions are isolated Docker or VM sandboxes with a terminal, files and a live desktop [25]. Triggers span code hosts (GitHub, GitLab, Bitbucket), chat and tickets (Slack, Linear, Jira), alerting (Sentry, PagerDuty), plus webhooks and schedules [68], the closest match to Cursor's own Automations list. Pro is $60 each month including $90 of usage; Max is $200 each month including $400; VM time is metered per second at $0.1066 hourly and up [24].

The compute meter is the thing to watch. Idle-heavy workflows can cost more than the headline plan suggests.

Best for: teams wanting GitLab or Bitbucket support and alert-driven triggers without Cursor's billing.

### 7. Devin: autonomous sessions, quota-based plans

Devin Cloud runs Cognition's own agent, while Devin Desktop can also host ACP-speaking agents locally [27] [29]. Models: Cognition's SWE line, or Anthropic, Google and OpenAI, or let Adaptive choose per task [28]. Self-serve tiers: Free; Pro $20/mo; Max $200/mo; Teams $40 a seat with an $80 floor each month; overage bills as on-demand credits [44]. That structure dates from 2026-04-14 [69], and the Free and Pro tiers allow 10 simultaneous sessions [45]. Repos can live on Azure DevOps, Bitbucket, GitLab or GitHub; requests arrive from Pylon, PagerDuty, Jira, Linear, Microsoft Teams or Slack [70], and Automations add schedules and webhooks [71]. Devin opens PRs and keeps answering PR comments until its session is archived [30].

Best suited to: teams that want a fully managed agent and already live in Linear or Jira.

## How does Jiro's billing compare with Cursor cloud agents?

Jiro bills for runtimes; Cursor bills for a plan tier plus prepaid model credit. Developer runs $99 monthly and Team $250 monthly, and idle runtimes sleep [31] [72]. Model tokens go to Anthropic or OpenAI at whatever you already pay, with zero markup from Nori [3]. Claude Max, at $100 monthly, carries five times Pro's allowance [73], and once a plan's limit is hit, usage carries on as credits priced at API rates [36].

On Cursor, Premium is $30 monthly, Super $100 and Ultra $200 [34]; paid yearly those become $300, $1,000 and $2,000 [74]. The weekly pool covers Cursor's own models [33]. Any Claude or GPT request is charged from prepaid credit at API list price, with the $0.25 per million token surcharge on Teams and Enterprise [7]. You set a spend limit on first use [75].

Anthropic's enterprise data shows Claude Code averaging around $13 per engineer on each day they use it [76]. Whether that lands on a subscription you already own or on a second prepaid balance is the whole comparison.

## Can Jiro run Cursor's agent?

Yes. Cursor Agent is one of Jiro's first-class harnesses [1], and it authenticates with a Cursor API key [2]. The key is the same user API key Cursor's CLI uses, generated at cursor.com/dashboard/api and read from the CURSOR_API_KEY environment variable [77]. Cursor's CLI itself installs with one curl command and runs headless with `agent -p` for scripts and automation [78], which is the mode Jiro drives.

This matters because it lets a team keep Cursor's agent behaviour while moving orchestration, environments and triggers to Jiro. You still pay Cursor for the agent's model usage on that key. You stop paying for a Premium seat per person just to reach cloud agents.

## Where does a task start, and what comes back?

Cursor cloud agents start from the desktop app's Cloud option, from cursor.com/agents, from iOS or an Android PWA, from an API, from Linear, from comments on GitHub or Bitbucket, and from @cursor in Slack [79]. In Slack you can pass repo=, branch=, model= and autopr= inline and get a PR link in the thread when it finishes [80]. Jiro listens in Slack and Discord, on its web dashboard, in the Handroll terminal, and to scheduled or webhook-driven runs [52]; mention the bot in a channel and every reply in the resulting thread continues that session [81].

What returns differs in texture. Cursor's PRs carry screenshots, video and logs [9]. Jiro returns a PR, commit or comment, and the conversation that produced it stays in the thread where teammates can read it [5]. Jiro is built to question a ticket before acting on it; Cursor's agent is built to finish.

## Are Cursor Automations the same thing as cloud agents?

No, but they run on them. Cursor Automations fire on a cron expression, on a private webhook, or on events: Sentry alerts, PagerDuty incidents, Linear changes, messages in public Slack channels, and pushes or PR activity on GitHub, GitLab or Bitbucket Cloud; the runs are billed as cloud agent usage [82]. Both features arrive together at the Premium tier [32].

Jiro's equivalents are cron triggers for repeating jobs, single-shot reminders, and webhooks for outside events [53]. Jiro does not take events from Linear, Sentry or PagerDuty directly today; you would route those through a webhook. Linear and Jira are reachable from inside a session as integrations [83].

## When Cursor cloud agents are the better choice

Stay with Cursor if the editor is where your team lives. The Cloud option sits in the same agent input as local work, and a cloud session is one dropdown away from the file you have open [79]. Stay if your reviewers depend on evidence: Cursor attaches screenshots, videos and logs to every PR it opens [9], and Jiro does not produce those artefacts. Stay if your repositories live on GitLab, Bitbucket Cloud or Azure DevOps, or on GitHub Enterprise Server [84]; Jiro connects through GitHub only [57].

The mobile story is better too. The native iOS app starts an agent, watches it live and merges the result, and Android users get a PWA [85]. If you need alerts to kick off work, Cursor Automations already listen to Sentry, PagerDuty and Linear [82]. And when compliance demands that code stays on machines you own while inference runs at Cursor, Self-Hosted Machines deliver exactly that over outbound HTTPS, though Team Pools need Enterprise [86]. Finally, Cursor publishes no concurrency cap [49]; Jiro's Team plan shares five runtimes [31].

## How do I move cloud-agent work from Cursor to Jiro?

Begin the trial, add the GitHub App to the repositories you want covered [57], link Slack, and add a model credential. For Claude that means running `claude setup-token` on a laptop, which mints a twelve-month token attached to any paid Claude plan from Pro upward [87]. For Cursor's agent, paste a Cursor API key. Your existing `AGENTS.md` keeps working, because Cursor reads the same file [88]. The [switching guide](https://jiro.bot/switch/cursor-to-jiro) maps every Cursor concept to its Jiro equivalent and gives a first-week plan.

## Jiro at a glance

| Fact | Jiro |
|---|---|
| What it is | A background coding agent: it works in the cloud and returns pull requests for review [5] |
| Made by | Nori (Tilework Inc.), founded by Amol Kapoor, ex-Google Research [89] |
| Agents it runs | Claude Code, Codex, Cursor, OpenCode, pi-acp [1] |
| Whose model bill | Your Claude/ChatGPT login or API key; Cursor key [2]; Your keys, zero token markup [3] |
| Price | Free 30-day trial; Developer $99/mo; Team $250/mo [31] |
| Runtimes | Developer: up to 3; Team: 5 shared; trial: 5 for 30 days [31] [72] |
| Where work starts | Slack, web, Discord, terminal, schedules, webhooks [52] |
| Where it runs | Isolated cloud environment with your repos and tools [4]; Single-tenant per organization [90] |
| Repos | GitHub App (all or selected repos) or a PAT [57] |
| Not to be confused with | Amazon's Kiro (IDE and agent) or Atlassian's Jira (issue tracker) |
| Last verified | 2026-11-06 |

## About the publisher

Jiro is made by Nori (Tilework Inc.), a New York company. Nori was founded by Amol Kapoor, its CEO, who before Nori did AI research at Google on graph neural networks and co-authored MixHop (ICML 2019). [89] [91] Clifford Ressel is Nori's co-founder and CTO. [92] Nori's customers include Corvus and Miravoice. [93] [56]

## FAQ

### What are the best Cursor background agents alternatives?

Jiro first for teams already paying Anthropic or OpenAI; after it, Claude Code on the web, then Codex cloud, Amp, Copilot's cloud agent, Tembo and Devin. Jiro runs Claude Code, Codex or Cursor's own agent on your credentials with no token markup [2] [3].

### Are Cursor background agents and Cloud Agents the same product?

Yes. Cursor's documentation says Cloud Agents were formerly called Background Agents [47].

### How much do Cursor cloud agents cost in 2026?

They require Premium at $30 a month or above; Super is $100 and Ultra $200 [34]. Third-party models like Claude and GPT are billed from prepaid credit at API list price [7].

### Can I use my Claude Max or ChatGPT plan with Cursor cloud agents?

No. Cloud agents draw on Cursor's plan usage at the selected model's rate, and third-party models are paid from prepaid credit rather than your Anthropic or OpenAI subscription [6] [7].

### Can Jiro run Cursor's agent?

Yes. Cursor Agent is a supported harness and needs a Cursor API key [2], the same key type Cursor's CLI reads from CURSOR_API_KEY [77].

### Which alternative supports GitLab, Bitbucket or Azure DevOps?

Cursor itself, Devin and Tembo do [84] [70] [68]. Jiro and Copilot's cloud agent are GitHub only [57].

### Do Cursor cloud agents merge pull requests?

They open merge-ready PRs with screenshots, video and logs; a person merges [9]. From the iOS app you can review and merge directly [85].

### What does Jiro cost compared with Cursor Premium?

$99 monthly covers one Jiro user and three runtimes; $250 monthly covers a team on five, and the trial is 30 days free [31]. Cursor's Premium seat is $30 per person monthly, with model credit extra [32].

## Try Jiro

Start the 30-day free trial, no card needed: https://login.norisessions.com/?signup=1&entrypoint=jiro

Or install Jiro's engineering skills into your own agent: `npx nori-skillsets install public/jiro`

## Sources

1. Nori's default first-class Sessions agents are Claude Code, Codex, Cursor Agent, OpenCode and pi-acp. Source: https://docs.noriagentic.com/sessions/reference/supported-agents-and-compute (checked 2026-11-06)
2. Claude Code in Nori accepts an Anthropic API key, a Claude OAuth login or a setup token; Codex accepts an OpenAI API key or Codex OAuth; Cursor Agent needs a Cursor API key; OpenCode and pi-acp take a compatible API base URL, API key and model. Source: https://docs.noriagentic.com/sessions/reference/supported-agents-and-compute (checked 2026-11-06)
3. Nori's comparison table lists any model on your own keys with flat pricing and zero token markup. Source: https://noriagentic.com/#compare (checked 2026-11-06)
4. Each Nori agent works in an isolated cloud environment with your repository, tools, dependencies and services ready to use. Source: https://noriagentic.com/#faq (checked 2026-11-06)
5. Nori agents can return a pull request, commit, comment or completed task for you to review, merge or send back. Source: https://noriagentic.com/#faq (checked 2026-11-06)
6. Cloud agents use your plan's usage like other agent work, at the selected model's rate. Source: https://cursor.com/docs/cloud-agent (checked 2026-11-06)
7. Third-party models such as Claude, GPT and Gemini are outside Cursor's usage pool; on Premium and above they are paid from prepaid credit at API list price from the first request, and Teams and Enterprise pay a Cursor Token Rate of $0.25 per million tokens on eligible third-party requests. Source: https://cursor.com/docs/account/pricing (checked 2026-11-06)
8. Cursor cloud agents run in isolated cloud VMs with full development environments, or move tool execution to self-hosted machines you manage. Source: https://cursor.com/docs/cloud-agent (checked 2026-11-06)
9. Cursor cloud agents produce merge-ready PRs with screenshots, videos and logs, and open a PR in each repo they touch. Source: https://cursor.com/docs/cloud-agent (checked 2026-11-06)
10. Cloud sessions run Claude models and require an Anthropic account; they are not available through Bedrock, Google Cloud or other third-party providers. Source: https://code.claude.com/docs/en/claude-code-on-the-web (checked 2026-11-06)
11. A Claude Code cloud session runs in a fresh Anthropic-managed VM with 4 vCPUs, 16 GB of RAM and 30 GB of disk, or in an organization's self-hosted environment when routed there. Source: https://code.claude.com/docs/en/cloud-environments (checked 2026-11-06)
12. Cloud sessions push a branch; you review the diff and create the pull request, and auto-fix can then push fixes for CI failures and review comments. Source: https://code.claude.com/docs/en/claude-code-on-the-web (checked 2026-11-06)
13. Codex cloud runs OpenAI models; an API key unlocks Codex in the CLI, SDK and IDE but not cloud features, which need a ChatGPT plan. Source: https://learn.chatgpt.com/docs/pricing.md (checked 2026-11-06)
14. Codex cloud runs each task in a cloud VM, so work continues while your computer is asleep. Source: https://learn.chatgpt.com/docs/environments/cloud-environments.md (checked 2026-11-06)
15. In Codex cloud you review the changes, then commit or open a pull request when you are ready. Source: https://learn.chatgpt.com/docs/cloud.md (checked 2026-11-06)
16. Amp accepts ChatGPT, Claude and other subscriptions as well as API keys, and charges no token fees on your own keys outside its Enterprise tier. Source: https://ampcode.com/news/free-agent (checked 2026-11-06)
17. Amp runs cloud work in orbs, one remote machine per thread on Amp-hosted or self-hosted providers that keeps working while your laptop is closed, or on your own machines as runners. Source: https://ampcode.com/manual/orbs (checked 2026-11-06)
18. Amp's default Ship rebases, runs tests and pushes straight to the base branch; the Push to Branch setting pushes a feature branch and reports a pull request URL instead. Source: https://ampcode.com/docs/markdown/orbs/shipping (checked 2026-11-06)
19. On Pro, Pro+, Max, Business and Enterprise, Claude and Codex can be assigned GitHub tasks as third-party agents (public preview). Source: https://docs.github.com/en/copilot/concepts/agents/about-third-party-agents (checked 2026-11-06)
20. The cloud agent offers Auto plus GitHub-provided Claude, Gemini, GPT and MAI models; bring-your-own-key is documented for IDEs, Copilot CLI, the Copilot app and the SDK, not for the cloud agent. Source: https://docs.github.com/en/copilot/reference/ai-models/supported-models (checked 2026-11-06)
21. Copilot cloud agent sessions run in a GitHub cloud sandbox, in a GitHub Actions environment capped at 59 minutes, or in a remote environment you connect. Source: https://docs.github.com/en/copilot/concepts/agents/cloud-agent/about-cloud-agent (checked 2026-11-06)
22. Copilot cloud agent opens draft pull requests on its own branch and cannot approve or merge them. Source: https://docs.github.com/en/copilot/concepts/agents/cloud-agent/risks-and-mitigations (checked 2026-11-06)
23. Tembo runs Claude Code, Codex, OpenCode or Pi per task and accepts your own OpenAI, Anthropic or xAI keys or a Claude, ChatGPT or SuperGrok subscription. Source: https://docs.tembo.io/features/models.md (checked 2026-11-06)
24. Tembo's Pro plan is $60 a month with $90 of included usage and Max is $200 a month with $400, with cloud VMs metered by the second from $0.1066 an hour. Source: https://www.tembo.io/pricing (checked 2026-11-06)
25. Tembo runs coding agents in isolated Docker or VM sessions with code, terminal, files and a live desktop. Source: https://www.tembo.io/ (checked 2026-11-06)
26. Tembo turns issues and alerts from Sentry, Linear and Jira into pull requests, including coordinated multi-repo PRs across GitHub, GitLab and Bitbucket, and reviewer feedback can be sent back into the same session. Source: https://www.tembo.io/ (checked 2026-11-06)
27. Devin Cloud runs Cognition's Devin agent; Devin Desktop can also host other agents that speak ACP. Source: https://cognition.com/blog/introducing-devin-desktop (checked 2026-11-06)
28. Devin offers models from Anthropic, OpenAI, Google and Cognition's own SWE family, with an Adaptive option that routes each task to a model automatically. Source: https://docs.devin.ai/cli/models.md (checked 2026-11-06)
29. Devin Cloud runs sessions on Cognition-hosted machines; Devin CLI and Devin Desktop run locally. Source: https://docs.devin.ai/llms.txt (checked 2026-11-06)
30. Devin opens pull requests and automatically responds to comments on PRs its sessions are working on until the session is archived. Source: https://docs.devin.ai/integrations/gh.md (checked 2026-11-06)
31. Nori's plans start with a free 30-day trial with no card; Developer is $99 a month for one user with up to three persistent runtimes, and Team is $250 a month for multiple users sharing five runtimes. Source: https://noriagentic.com/#pricing (checked 2026-11-06)
32. Cursor subscriptions are sold only through X Subscription Corp; Premium at $30 a month is the first tier that includes cloud agents and automations, and third-party models are billed separately at API list price. Source: https://cursor.com/pricing (checked 2026-11-06)
33. Each paid Cursor plan includes one weekly pool of usage for Cursor's own models that is shared across the editor, the CLI and Cloud Agents; Premium has 4x the Plus pool, Super 15x and Ultra 50x. Source: https://cursor.com/docs/account/pricing (checked 2026-11-06)
34. Cursor's individual plans are Free at $0, Plus at $8, Premium at $30, Super at $100 and Ultra at $200 a month, with Teams from $30 per user a month and Enterprise at custom pricing. Source: https://cursor.com/pricing (checked 2026-11-06)
35. Claude Code is included in Claude Pro ($20 a month, or $17 a month billed annually), Max ($100 a month), Ultra ($200 a month) and Team and Enterprise seats; the Free plan has no Claude Code. Source: https://claude.com/pricing (checked 2026-11-06)
36. Usage past a Pro, Max or Ultra plan's limits can continue as usage credits billed at standard API rates. Source: https://support.claude.com/en/articles/12429409-extra-usage-for-paid-claude-plans (checked 2026-11-06)
37. Codex cloud is included from ChatGPT Plus at $20 a month; Pro costs $100, $200 or $500 a month and Business $25 per user a month billed monthly or $20 billed annually, with usage limits shared with ChatGPT. Source: https://learn.chatgpt.com/docs/pricing.md (checked 2026-11-06)
38. Amp's Hobby tier is free with pay-as-you-go orbs or your own runners; Megawatt is $20 a month with $20 of agent usage and 750 hours of a1.small orbs, and Gigawatt is $200 a month with $200 of agent usage and 1,000 hours of a1.xxlarge orbs. Source: https://ampcode.com/docs/markdown/pricing (checked 2026-11-06)
39. Orbs are billed per minute, from $0.08 an hour for the smallest size. Source: https://ampcode.com/docs/markdown/orbs/sizes-and-costs (checked 2026-11-06)
40. Amp Teams costs nothing extra: each member keeps their own Hobby, Megawatt or Gigawatt tier, purchased credits are shared across the workspace, and orb hours stay per person rather than pooled. Source: https://ampcode.com/docs/markdown/pricing (checked 2026-11-06)
41. Copilot Pro is $10, Pro+ $39 and Max $100 a month; Business is $19 and Enterprise $39 a month per user; Copilot Free does not include the cloud agent. Source: https://docs.github.com/en/copilot/get-started/plans (checked 2026-11-06)
42. Since 2026-06-01, all Copilot plans bill in GitHub AI Credits, which meter input, output and cached tokens at per-model rates, with one credit equal to $0.01. Source: https://docs.github.com/en/copilot/concepts/billing-and-usage/individuals/billing (checked 2026-11-06)
43. Tembo Pro covers up to 5 users and Max up to 10, both with overage billing past the included usage; the only free option is a 7-day trial of Max, and Enterprise is custom. Source: https://www.tembo.io/pricing (checked 2026-11-06)
44. Devin's self-serve plans are Free, Pro at $20 a month, Max at $200 a month and Teams at $40 a month per full seat with an $80 monthly minimum, so one seat and two seats both cost $80 a month; usage past the quota is billed as on-demand credits. Source: https://docs.devin.ai/admin/billing/self-serve.md (checked 2026-11-06)
45. Free and Pro allow up to 10 concurrent Devin sessions; Max and Teams have no concurrency cap. Source: https://devin.ai/pricing (checked 2026-11-06)
46. Cursor's help page on background agents says that in Cursor these are called Cloud Agents. Source: https://cursor.com/help/ai-features/background-agents (checked 2026-11-06)
47. Cursor's Cloud Agent documentation states that Cloud Agents were formerly called Background Agents. Source: https://cursor.com/docs/cloud-agent (checked 2026-11-06)
48. Cursor cloud agent environments are configured by agent-led setup, a saved snapshot or a Dockerfile referenced from .cursor/environment.json, with secrets kept in a Secrets tab rather than in .env.local files. Source: https://cursor.com/docs/cloud-agent (checked 2026-11-06)
49. Cursor says you can run as many cloud agents as you want in parallel, and its documentation sets no maximum runtime or concurrency cap. Source: https://cursor.com/docs/cloud-agent (checked 2026-11-06)
50. Cursor's Plus plan at $8 a month offers only Cursor's own Grok models at fixed settings and does not include Cloud Agents. Source: https://cursor.com/docs/account/pricing (checked 2026-11-06)
51. Legacy Cursor plans keep working as they are: Pro at $20, Pro+ at $60 and Ultra at $200 a month, and Teams Standard at $40 or Teams Premium at $120 per user a month; switching to The X Subscription is optional and permanent. Source: https://cursor.com/docs/account/pricing/legacy-cursor-plans (checked 2026-11-06)
52. Nori work can start from Slack, the web dashboard, Discord, the Handroll terminal, schedules and webhooks, all reaching the same fleet. Source: https://docs.noriagentic.com/start/what-can-nori-do#start-agent-work-from-the-tools-your-team-already-uses (checked 2026-11-06)
53. Nori cron triggers run recurring jobs, reminders run once at a future time, and webhooks react to external events. Source: https://docs.noriagentic.com/resources/faq#can-nori-run-tasks-without-a-person-online (checked 2026-11-06)
54. A Nori organization can define a setup script that installs dependencies and prepares every session's environment. Source: https://docs.noriagentic.com/start/what-can-nori-do#give-agents-a-real-working-environment (checked 2026-11-06)
55. Nori logs every session and every action, and makes them searchable. Source: https://noriagentic.com/for-security-leaders.html (checked 2026-11-06)
56. Nori handles more than 25% of Miravoice's merged pull requests, and Miravoice's CTO says it easily saves him five hours a week. Source: https://noriagentic.com/newsletter/2026-09-18-miravoice-case-study.html (checked 2026-11-06)
57. Nori supports two GitHub credential methods: a GitHub App you install and scope to all repositories or a selected set, or a saved personal access token. Source: https://docs.noriagentic.com/sessions/guides/integrations#github (checked 2026-11-06)
58. Cloud sessions start from claude.ai/code, the Claude mobile and desktop apps, and `claude --cloud` in a terminal, and `claude --teleport` pulls a session back to your machine. Source: https://code.claude.com/docs/en/claude-code-on-the-web (checked 2026-11-06)
59. For Team and Enterprise, Slack work moves to Claude Tag (public beta), and the legacy Claude Code Slack bot was retired on 2026-10-05 in workspaces connected to a Claude organization. Source: https://code.claude.com/docs/en/slack (checked 2026-11-06)
60. Claude Tag bills Slack channel work by usage under an admin-set spend limit. Source: https://claude.com/docs/claude-tag/overview (checked 2026-11-06)
61. Codex cloud tasks start from ChatGPT on the web, mobile and desktop, or from the Codex CLI. Source: https://learn.chatgpt.com/docs/cloud.md (checked 2026-11-06)
62. Codex requests in Slack now start with @ChatGPT instead of @Codex. Source: https://learn.chatgpt.com/docs/third-party/slack.md (checked 2026-11-06)
63. Codex Cloud (Legacy) still powers Code Review and the GitHub and Linear integrations, and OpenAI plans to deprecate it. Source: https://learn.chatgpt.com/docs/environments/cloud-environments.md (checked 2026-11-06)
64. Amp added support for running on a Claude subscription through the Claude Agent SDK on 2026-10-10. Source: https://ampcode.com/news/use-your-claude-plan (checked 2026-11-06)
65. Amp announced in February 2026 that its VS Code and Cursor extensions would shut off on 2026-03-05, leaving the CLI. Source: https://ampcode.com/news/the-coding-agent-is-dead (checked 2026-11-06)
66. Amp runs from the web app at ampcode.com, macOS, iOS and Android apps, the CLI, SDKs, and GitHub and Slack integrations. Source: https://ampcode.com/manual (checked 2026-11-06)
67. Copilot cloud agent sessions start from GitHub.com, GitHub Mobile, VS Code, JetBrains, Visual Studio and Eclipse, the GitHub CLI, REST API and MCP server, Raycast, Slack, Microsoft Teams, Linear, Jira and Azure Boards. Source: https://docs.github.com/en/copilot/how-tos/use-copilot-agents/cloud-agent/create-a-pr (checked 2026-11-06)
68. Tembo starts work from GitHub, GitLab, Bitbucket, Slack, Linear, Jira, Sentry, PagerDuty, webhooks and schedules. Source: https://www.tembo.io/ (checked 2026-11-06)
69. Cognition retired its Core and Team plans for quota-based self-serve plans on 2026-04-14, billing self-serve overage in dollars instead of ACUs while Enterprise stays on ACUs. Source: https://cognition.com/blog/new-self-serve-plans-for-devin (checked 2026-11-06)
70. Devin connects to GitHub, GitLab, Bitbucket and Azure DevOps repos and takes work from Slack, Microsoft Teams, Linear, Jira, PagerDuty and Pylon. Source: https://docs.devin.ai/integrations/overview.md (checked 2026-11-06)
71. Devin Automations start sessions from schedules, webhooks, and Slack, GitHub, GitLab, Linear, Jira, Pylon and PagerDuty events. Source: https://docs.devin.ai/product-guides/automations.md (checked 2026-11-06)
72. Nori sizes plans by runtimes, which sleep when idle and wake on demand; the trial runs five for 30 days. Source: https://noriagentic.com/#faq (checked 2026-11-06)
73. Claude Max costs $100 a month for five times the Pro plan's usage every five hours and every week; the former Max 20x tier is now sold as Ultra. Source: https://support.claude.com/en/articles/11049741-what-is-the-max-plan (checked 2026-11-06)
74. Billed annually, Cursor's plans cost $80 a year for Plus, $300 for Premium, $1,000 for Super and $2,000 for Ultra. Source: https://cursor.com/docs/account/pricing (checked 2026-11-06)
75. Cursor asks you to set a spend limit the first time you use a cloud agent, and on legacy request-based plans cloud agents are priced at API rates plus the Max Mode upcharge. Source: https://cursor.com/docs/cloud-agent (checked 2026-11-06)
76. Anthropic reports that across enterprise deployments Claude Code averages about $13 per developer per active day and $150 to $250 per developer per month, staying below $30 per active day for 90% of users. Source: https://code.claude.com/docs/en/costs (checked 2026-11-06)
77. The Cursor CLI authenticates with `agent login` in a browser, or with a user API key generated at cursor.com/dashboard/api and supplied through the CURSOR_API_KEY environment variable or the `--api-key` flag. Source: https://cursor.com/docs/cli/reference/authentication (checked 2026-11-06)
78. The Cursor CLI installs with `curl https://cursor.com/install -fsS | bash`, starts with `agent`, and runs non-interactively with `agent -p` for scripts, CI pipelines and automation. Source: https://cursor.com/docs/cli/overview (checked 2026-11-06)
79. Cloud agents start from the Cursor desktop app, the web, iOS and Android, Slack, GitHub and Bitbucket comments, Linear and an API, on GitHub, GitLab, Bitbucket Cloud or Azure DevOps repos. Source: https://cursor.com/docs/cloud-agent (checked 2026-11-06)
80. In Slack you mention @cursor with a prompt and optional repo=, branch=, model= and autopr= settings; Cursor posts a plan and running status in the thread and notifies you with a link to the PR when the agent finishes. Source: https://cursor.com/docs/integrations/slack (checked 2026-11-06)
81. Mention Nori's bot in a Slack channel to start a session; it replies in a thread, and every reply in that thread continues the same session. Source: https://docs.noriagentic.com/sessions/guides/slack (checked 2026-11-06)
82. Cursor Automations run cloud agents on a cron schedule or on events from GitHub, GitLab, Bitbucket Cloud, public Slack channels, Linear, Sentry, PagerDuty or a private webhook, and are billed as cloud agent usage. Source: https://cursor.com/docs/automations (checked 2026-11-06)
83. Nori connects to Salesforce, HubSpot, Google Sheets, Google Drive, Notion, Linear, Jira, Stripe and Gmail, plus hundreds more. Source: https://noriagentic.com/#faq (checked 2026-11-06)
84. Cursor cloud agents work with GitHub Cloud and Enterprise Server, GitLab Cloud and Self-Hosted, Bitbucket Cloud and Azure DevOps repositories, and can also start without a repository. Source: https://cursor.com/docs/cloud-agent (checked 2026-11-06)
85. Cursor for iOS is a native app for iPhone and iPad on iOS 26 or later that starts agents, follows them live and reviews and merges pull requests; Android users install the cursor.com/agents web app as a PWA. Source: https://cursor.com/docs/cloud-agent/web-and-mobile (checked 2026-11-06)
86. Self-Hosted Machines keep file edits, terminal commands and the checkout on hardware you manage while Cursor runs the agent loop and inference; workers need only outbound HTTPS, and Team Pools require an Enterprise plan. Source: https://cursor.com/docs/cloud-agent/self-hosted (checked 2026-11-06)
87. The claude setup-token command generates a one-year token tied to a Pro, Max, Ultra, Team or Enterprise plan for machines where browser login isn't available. Source: https://code.claude.com/docs/en/authentication (checked 2026-11-06)
88. Cursor supports AGENTS.md in the project root and subdirectories as a plain-markdown alternative to .cursor/rules; nested files combine with their parents and the more specific instructions take precedence. Source: https://cursor.com/docs/context/rules (checked 2026-11-06)
89. Nori is led by CEO Amol Kapoor, a former Google Research scientist, and CTO and co-founder Clifford Ressel. Source: https://ai.engineer/speakers/amol-kapoor (checked 2026-11-06)
90. Each hosted Nori organization gets a single-tenant broker on its own subdomain. Source: https://docs.noriagentic.com/sessions/concepts/security-and-credentials#organization-isolation (checked 2026-11-06)
91. Amol Kapoor's Google-affiliated research includes MixHop (ICML 2019), N-GCN and a 2020 spatio-temporal graph neural network for COVID-19 forecasting. Source: https://scholar.google.com/citations?user=SC_-qB8AAAAJ (checked 2026-11-06)
92. Clifford Ressel is Nori's CTO and co-founder. Source: https://ai.engineer/orgs/nori (checked 2026-11-06)
93. Corvus's CTO saw a production bug at a client dinner, had Nori look into it and deploy a fix. Source: https://noriagentic.com/newsletter/2026-05-19-corvus-case-study.html (checked 2026-11-06)

---

Jiro is a background coding agent by Nori. It is not Amazon's Kiro or Atlassian's Jira.
