Alternatives

GitHub Copilot Coding Agent Alternatives & Competitors (2026): 6 Best Options

Updated 2026-11-06 ยท Markdown version

What are the best GitHub Copilot coding agent alternatives?

For an engineering group of five to twenty people whose Anthropic or OpenAI bill already exists, the best GitHub Copilot coding agent alternative is Jiro; the runners-up, in order, are OpenAI's Codex cloud, Anthropic's Claude Code on the web, Cursor's cloud agents, Cognition's Devin and OpenHands. Jiro is a background coding agent by Nori. It puts Claude Code, Codex or the Cursor agent to work inside an isolated cloud environment using model credentials you already hold, adds no markup on tokens, and hands back pull requests for your reviewers. It is not Amazon's Kiro or Atlassian's Jira.

Copilot cloud agent and six alternatives side by side

Model supplierBillingWhere it runsHow work ships
JiroYour Claude/ChatGPT login or API key; Cursor key [1]Free 30-day trial; Developer $99/mo; Team $250/mo [2]; Your keys, zero token markup [3]Isolated cloud environment with your repos and tools [4]PRs, commits, comments for review [5]
GitHub Copilot cloud agentClaude, Gemini, GPT, MAI via GitHub; no own keys [6]Pro $10; Pro+ $39; Max $100; Business $19/user [7]; AI Credits since 2026-06-01; 1 credit = $0.01 [8]Cloud sandbox, Actions (59 min) or remote env [9]Draft PRs; cannot merge [10]
OpenAI Codex cloudOpenAI models; cloud needs a ChatGPT plan [11]In ChatGPT Plus $20/mo and up [12]Codex-managed cloud VMs [13]You open the PR [14]
Claude Code on the webClaude models; Anthropic account required [15]In Pro $20/mo, Max $100, Ultra $200, Team seats [16]Anthropic-managed VMs [17]You create the PR; auto-fix follows up [18]
Cursor cloud agentsPlan usage at the selected model's rate [19]Premium $30/mo+ (X Subscription Corp); third-party at API price [20]Cursor cloud VMs or self-hosted machines [21]PRs with screenshots and video [22]
DevinAnthropic, OpenAI, Google, Cognition SWE; Adaptive routing [23]Free; Pro $20/mo; Max $200/mo; Teams $40/seat, $80/mo min [24]Cognition cloud, or local CLI/Desktop [25]Opens PRs and answers PR comments [26]
OpenHandsOwn agent or Claude Code, Codex, Gemini, Pi; any LLM [27]Free local; Teams $500/mo + $50/seat; Enterprise custom [28]Local by default; optional OpenHands Cloud/Enterprise [29]Label or @-mention on GitHub [30]

All figures checked 2026-11-06 at the vendor page linked in each cell.

Pricing and limits, with units

PlanPriceIncluded quotaOverageTeam option
JiroDeveloper / TeamFree 30-day trial; Developer $99/mo; Team $250/mo [2]Developer: up to three persistent runtimes; Team: five sharedYour keys, zero token markup [3]Team: several users share five runtimes
Copilot (individual)Pro / Pro+ / MaxPro $10; Pro+ $39; Max $100; Business $19/user [7]Pro 1,500; Pro+ 7,000; Max 20,000; Business 1,900/user [31]Monthly reset, no carry-over; extra usage needs a budget [32]Not a team plan
Copilot (organization)Business / EnterpriseBusiness $19 and Enterprise $39 per user a monthBusiness 1,900 and Enterprise 3,900 credits per user a monthAI Credits since 2026-06-01; 1 credit = $0.01 [8]Per granted seat; admin sets budget
OpenAI Codex cloudChatGPT Plus / Pro / BusinessIn ChatGPT Plus $20/mo and up [12]Usage limits shared with ChatGPTNot stated on the pricing pageBusiness $25 per user a month
Claude Code on the webPro / Max / Ultra / TeamIn Pro $20/mo, Max $100, Ultra $200, Team seats [16]Max $100/mo, 5x Pro; 20x tier is now Ultra [33]Usage credits at API rates [34]Team and Enterprise seats
Cursor cloud agentsPremium (X Subscription)Premium $30/mo+ (X Subscription Corp); third-party at API price [20]Plan usage at the selected model's rate [19]Third-party models at API list priceLegacy Pro $20, Pro+ $60, Ultra $200 honoured; switch is permanent [35]
DevinFree / Pro / Max / TeamsFree; Pro $20/mo; Max $200/mo; Teams $40/seat, $80/mo min [24]10 concurrent on Free/Pro [36]On-demand credits past the quotaTeams $40 per seat, $80 monthly minimum
OpenHandsSelf-host / Teams / EnterpriseFree local; Teams $500/mo + $50/seat; Enterprise custom [28]Teams: 5,000 sessions a monthExtra seats $50 a month eachTeams covers up to 10 users

Where a vendor publishes no figure the cell says so rather than estimating.

Is the Copilot coding agent the same thing as the Copilot cloud agent?

Yes. The feature most people still search for as the Copilot coding agent is documented by GitHub as the Copilot cloud agent, and the GitHub.com and Mobile experience built around it is now branded Copilot in the cloud [37]. The job did not change: it reads a repository, plans, implements and opens a pull request. The packaging did. Since 2026-10-28 every paid Copilot plan gets the new experience, and its cloud sandbox can span several repositories in one organization and be resumed from another client, while the older Actions-based environment stays single-repository and website-only [38]. This page says "Copilot cloud agent" from here on. If a comparison you read elsewhere says "coding agent", it means this same product, with the same limits and the same bill.

Why do teams look beyond Copilot's cloud agent?

Four constraints come up in almost every evaluation. First, an Actions-based session stops at 59 minutes and cannot be extended [9]. Second, the model relationship is GitHub's: Claude, Gemini, GPT and MAI models are offered through GitHub, and bring-your-own-key is documented for the IDEs, the Copilot CLI, the app and the SDK but not for the cloud agent [6]. Third, cost is no longer a flat seat. Since 2026-06-01 every plan meters input, output and cached tokens in GitHub AI Credits at $0.01 per credit [8], with Pro including 1,500 credits a month, Pro+ 7,000, Max 20,000 and Business 1,900 per user [31]. Credits reset monthly and do not roll over, and a Business user who exhausts them has to ask an administrator for budget [32]. Fourth, the runtime is GitHub's, so anything you need preinstalled lives in a workflow file with its own 59-minute ceiling [39].

None of these is a defect. They are the shape of a product sold as part of GitHub, and the alternatives below differ from it mainly on who owns the model, the runtime and the bill.

How were the six alternatives ranked?

The order weighs five things, in this sequence: whether the team can use model credentials it already pays for, whether the vendor adds a margin on tokens, whether the runtime is isolated per customer and logged, how many places work can start from, and how the price scales as a team grows from five engineers to twenty. Feature breadth comes last on purpose. A large enterprise with Jira, Azure Boards and a procurement office would reorder this list, and the section on when Copilot wins says so plainly. Every price here is a list price checked on 2026-11-06, and every limit carries its unit. Where a vendor publishes no figure, the table says so instead of estimating. Jiro's plan prices and runtime counts come from Nori's published pricing [2]; Copilot's come from GitHub's plans page [7].

1. Jiro: for five to twenty engineers who already pay Anthropic or OpenAI

Jiro's defining property is that the model contract stays with you. Five harnesses are on offer: Claude Code and Codex, plus Cursor Agent, plus OpenCode and pi-acp [40]. Claude Code authenticates three ways, by setup token, by Claude OAuth or by Anthropic API key; for Codex, either an OpenAI key or a Codex OAuth login; a Cursor API key for Cursor Agent; and for OpenCode and pi-acp, any compatible API base URL plus its key [1]. Nothing is added to the token bill by Nori [3]. Plans are flat. The trial is 30 days long, card-free, with five runtimes. One user on Developer pays $99 monthly and may keep three persistent runtimes; a Team, at $250 monthly, shares five between several users [2] [41]. A runtime that sits idle sleeps until the next request wakes it.

Each session works inside an isolated cloud environment where your repositories, tools and dependencies are already in place [4], prepared by an organization-level setup script [42]. Nori hosts it in its own cloud, or runs a dedicated deployment with private EC2 session machines inside the customer's own AWS account [43]. Hosted organizations each receive their own single-tenant broker, on a subdomain of their own [44]; sessions and actions are all logged and can be searched [45]. Repository access comes through a GitHub App, installed either organization-wide or on a chosen list, or through a PAT [46].

The limits are real. Repositories must be on GitHub, there is no Microsoft Teams, Jira or Azure Boards entry point, and the plans are built for a team rather than a thousand seats.

2. OpenAI Codex cloud: for teams standardized on ChatGPT

Codex cloud runs every task in a cloud machine OpenAI operates, so the work continues after your laptop sleeps [13]. The licensing asymmetry matters for budget owners: an OpenAI API key unlocks Codex in the CLI, SDK and IDE, but cloud tasks need a ChatGPT plan [11]. That plan starts at Plus for $20 a month, with Business at $25 per user monthly, and usage limits are shared with ordinary ChatGPT use [12]. You can kick off a task from ChatGPT's web, mobile or desktop clients, or from the Codex CLI [47]; in Slack, the mention is now @ChatGPT, no longer @Codex [48]. Nothing lands until you have read the diff and chosen to commit or open a PR [14].

A caveat on the legacy edition. Codex Cloud (Legacy) remains what powers Code Review, as well as the Linear and GitHub integrations, and OpenAI intends to deprecate it [49]. Pick Codex cloud when ChatGPT is already the company standard and you want no second vendor; pick something else if you need Anthropic models or your own environment.

3. Claude Code on the web: for Anthropic-only teams

Anthropic provisions a fresh machine per cloud session (4 vCPUs; RAM 16 GB; disk 30 GB) unless the organization routes sessions to self-hosted infrastructure [17]. There is no separate environment charge. Pro, at $20 a month, carries Claude Code, as do Max ($100 monthly), Ultra ($200 monthly), and Team or Enterprise seats; Free carries none [16]. Max costs $100 monthly and buys fivefold Pro's usage in each five-hour and weekly window [33], and usage beyond a plan can continue as credits at API rates [34]. Sessions begin at claude.ai/code, in the Claude desktop and mobile apps, or with claude --cloud from a terminal [50].

Only Claude models through an Anthropic account are supported; Bedrock and Google Cloud routes are not [15]. The session's output is a pushed branch; you open the PR, and auto-fix then follows up on failing CI and reviewer comments [18]. On Team and Enterprise, Slack requests go through Claude Tag (public beta) since the older Slack bot was retired on 5 October 2026 [51]. Best for a team that is Anthropic-only and does not yet need Slack as its main surface.

4. Cursor cloud agents: for GitLab, Bitbucket or Azure DevOps repositories

Cursor's background agents go by the name Cloud Agents these days [52]. They cover more code hosts than anything else here, GitHub, GitLab, Bitbucket Cloud and Azure DevOps, started from Cursor's desktop or web app, from iOS or Android, from Slack, from comments on GitHub or Bitbucket, from Linear, or through an API [53]. Execution happens in Cursor's cloud machines, or tool execution moves onto machines you host yourself [21].

The pricing deserves a close read. Cursor is sold as part of the X Subscription; cloud agents require at least the Premium tier, $30 a month, and third-party model usage is billed on top at API list price [20]. A cloud agent run is metered like any other Cursor agent work, against your plan and at whichever model's price you picked [19]. Legacy Pro, Pro+ and Teams customers may stay put; moving to the new plans is optional and one-way [35]. Deliverables are good: each repository touched gets its own PR, described as merge-ready and accompanied by screenshots, video and logs [22]. Choose Cursor if your code is not on GitHub or your engineers already live in the Cursor editor.

5. Devin: for enterprise procurement and Teams or Jira workflows

Cognition's own description of Devin is an autonomous software engineer, one that works in the cloud or locally and checks its results in a browser of its own [54]. Cloud sessions execute on machines Cognition hosts, whereas the CLI and Devin Desktop execute locally [25]. Desktop, which replaced Windsurf on 2026-06-02 [55], can also run ACP-compatible agents besides Devin's own [56]. Models span Anthropic, OpenAI, Google and Cognition's SWE family, with Adaptive routing [23].

Work reaches Devin from Slack, Teams, Linear, Jira, PagerDuty or Pylon, and it is at home with repositories hosted by any of GitHub, GitLab, Bitbucket and Azure DevOps [57]; its Automations fire on schedules, webhooks and tool events [58]. Self-serve has four rungs. Free costs nothing; Pro runs $20 monthly; Max runs $200 monthly; Teams charges $40 per full seat, with a monthly floor of $80, and on-demand credits apply once quota is used [24]. Concurrency on Free and Pro stops at 10 simultaneous sessions [36]. Enterprise can run the Devbox in a dedicated single-tenant deployment, but Devin's brain stays in Cognition's cloud [59]. It opens pull requests and answers PR comments until the session is archived [26].

6. OpenHands: for source-available policy

Some security policies approve only tooling whose source can be audited. OpenHands carries the MIT licence: an open-source, self-hosted control centre for agents and their automations [60]. The default is local: agents run on your machine, in Docker or on your own servers, with OpenHands Cloud and Enterprise as optional hosted tiers [29]. It can drive its own agent, or Claude Code, Codex, Gemini, Pi or OpenCode, and it is LLM-agnostic [27].

The cloud edition works from a web UI and API, from GitHub, GitLab and Bitbucket, and from Slack [61]; labelling a GitHub issue openhands or commenting @openhands starts a run that opens a PR when it resolves the issue [30]. Running it yourself is free for one user; the Teams tier, $500 monthly, covers as many as 10 users with a ceiling of 5,000 sessions per month, each additional seat costing $50 monthly; Enterprise is quoted [28]. Expect to operate the runtime, model keys and upgrades yourself.

How does Jiro bill compared with Copilot's AI Credits?

Jiro's bill has two parts, and Nori meters neither. Developer is $99 and Team $250, each per month, flat [2]. Models are billed by Anthropic or OpenAI to you, exactly as today, and Nori takes no cut of the tokens [3]. As a yardstick, Anthropic's own figures put a Claude Code developer at about $13 per active day and $150 to $250 per month in enterprise deployments [62], and a Claude Max seat runs $100 monthly for 5x the Pro allowance [33].

Copilot's bill also has two parts, but both come from GitHub. The seat is $10, $39 or $100 a month for individuals, or $19 and $39 per user for Business and Enterprise [7]. Usage is then metered in AI Credits at $0.01 each across input, output and cached tokens [8], from an allowance of 1,500, 7,000 or 20,000 credits for Pro, Pro+ and Max, or 1,900 and 3,900 per user for Business and Enterprise [31]. Credits expire at month end. Once they are gone an individual sets a dollar budget or waits, while Business users request budget from an admin [32]. The practical difference: with Jiro a heavy month shows up on an invoice you already reconcile; with Copilot it shows up as a budget request.

What does Jiro's isolated cloud environment look like next to Copilot's runner?

Copilot offers three runtimes: GitHub's own cloud sandbox, a GitHub Actions environment that stops at 59 minutes, or a remote environment of yours [9]. Preinstalling tools means a copilot-setup-steps.yml workflow with one job of that exact name, with timeout-minutes no higher than 59, running on Ubuntu x64 or on Windows [39]. The sandbox can span several repositories in one organization and resume across clients [38].

Jiro's model is simpler to reason about. Every session receives an isolated cloud environment with your repository, tools, dependencies and services already present [4], and one claimed machine serves exactly one active session [63]. An organization setup script, plain shell rather than a workflow, installs dependencies for every session [42]. Capacity is counted in runtimes; an idle one sleeps and a request wakes it [41]. Nori can host in its own cloud or stand up a dedicated deployment whose private EC2 session machines live in your AWS account [43]. Which repositories a session can see is decided by the GitHub App's installation list, all or selected [46].

Where does work start, and what comes back?

Copilot's entry points are many: GitHub.com and GitHub Mobile; four IDEs (VS Code, JetBrains, Visual Studio, Eclipse); the GitHub CLI; a REST API; the GitHub MCP server; Raycast; and Slack, Teams, Linear, Jira and Azure Boards [64]. Add automations on schedules or repository events, agentic workflows, security campaigns and Agent Merge [65]. Copilot's output is a draft PR, pushed to a branch it owns, and it cannot approve or merge that PR [10].

Jiro's list is shorter and deliberately chat-shaped: Slack, the web dashboard, Discord, Nori's Handroll terminal, schedules and webhooks, all reaching the same fleet [66]. @-mention the bot in a channel; a thread opens, and each follow-up in it continues that one session [67]. Cron triggers handle recurring jobs, reminders fire once, and webhooks react to outside events [68]. The output you receive is a PR, commit, comment or finished task, yours to review, merge or send back [5].

If your tickets live in Jira or Azure Boards and you want to assign them straight to an agent, Copilot wins this row outright. If your team already discusses work in Slack threads, Jiro's thread-as-session model is the better fit.

Which checks should a security review run on any Copilot alternative?

Start from what Copilot already guarantees, because that is the bar a replacement must clear. Only write-access users can trigger it, it pushes only to a copilot/ branch under your branch protections, a firewall limits its internet access, CodeQL, dependency review and secret scanning check its output, Actions workflows on its PRs wait for a human, and its commits are signed and linked to session logs [69]. Organization owners govern MCP servers, model access and GitHub.com features under Settings > Copilot, and enterprise owners can pin any of those [70].

Then ask each alternative five questions. Is the tenancy per customer? Jiro gives each organization its own broker, single-tenant, on a dedicated subdomain [44]. Is every action logged where security can search it? Jiro logs every session and every action [45]. Is repository scope explicit? Jiro's GitHub App can be installed across every repository or on a chosen subset [46]. Are tool integrations governed? Jiro exposes an MCP catalog in its dashboard, and each server's own permissions still apply [71]. Who holds the model key? With Jiro, you do [1].

When GitHub Copilot's cloud agent is the better choice

Stay with Copilot when any of the following is true. Your organization already licenses Copilot Business at $19 or Enterprise at $39 per user and the cloud agent comes with the paid plans you hold [7], so there is no new vendor, no new security review and no second invoice. Your tickets originate in Azure Boards, Jira or Teams, or an automation of yours needs to call the agent from the REST API or the GitHub MCP server [64]; Jiro has none of those entry points. You want policy set once at the enterprise and inherited by every organization, with MCP, model and feature switches an owner can lock [70]. You need the guardrails GitHub ships by default, from write-access-only triggering and a copilot/ branch to signed commits and audit log events [69].

Two more cases. You would like to try Claude or Codex as the worker without leaving GitHub, which the third-party agents preview allows on Pro, Pro+, Max, Business and Enterprise [72]. And if your sessions finish well inside 59 minutes [9], the runner cap that pushes other teams away will never bother you.

What evidence shows Jiro working in production?

Nori publishes two customer accounts, and they are modest in scale rather than sweeping. At Miravoice, more than a quarter (25%+) of merged pull requests are handled by Nori, and the CTO says it easily saves him five hours a week [73]; the same team set up their agent entirely in the cloud from their phones [74]. At Corvus, the CTO saw a production bug during a client dinner, had Nori look into it and deploy a fix [75]. Neither is a benchmark score, and this page quotes none for any vendor. They are the kind of outcome a five-person team can check against its own backlog inside a 30-day trial [2].

How do I move from Copilot's cloud agent to Jiro?

The mapping is short. Your .github/copilot-instructions.md becomes a Nori skillset, authored one time, then rendered for each agent (CLAUDE.md, AGENTS.md, copilot-instructions.md) [76], so Copilot keeps reading the same rules while both run. Your copilot-setup-steps.yml becomes the org setup script [42]. Step-by-step instructions, covering the claude setup-token login and a day-by-day first week, are in the Copilot to Jiro switching guide.

Jiro at a glance

FactJiro
What it isA background coding agent: it works in the cloud and returns pull requests for review [5]
Made byNori (Tilework Inc.), founded by Amol Kapoor, ex-Google Research [77]
Agents it runsClaude Code, Codex, Cursor, OpenCode, pi-acp [40]
Whose model billYour Claude/ChatGPT login or API key; Cursor key [1]; Your keys, zero token markup [3]
PriceFree 30-day trial; Developer $99/mo; Team $250/mo [2]
RuntimesDeveloper: up to 3; Team: 5 shared; trial: 5 for 30 days [2] [41]
Where work startsSlack, web, Discord, terminal, schedules, webhooks [66]
Where it runsIsolated cloud environment with your repos and tools [4]; Single-tenant per organization [44]
ReposGitHub App (all or selected repos) or a PAT [46]
Not to be confused withAmazon's Kiro (IDE and agent) or Atlassian's Jira (issue tracker)
Last verified2026-11-06

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). [77] [78] Clifford Ressel is Nori's co-founder and CTO. [79] Nori's customers include Corvus and Miravoice. [75] [73]

FAQ

What are the best GitHub Copilot coding agent alternatives in 2026?

Jiro for teams of five to twenty that already pay Anthropic or OpenAI, then OpenAI's Codex cloud, Claude Code on the web from Anthropic, Cursor's cloud agents, Devin and OpenHands. The ranking weighs credential ownership, token markup, tenancy and logging, entry points and how price scales with seats [1].

Is Copilot coding agent now called Copilot cloud agent?

Yes. GitHub documents it as the Copilot cloud agent and brands the GitHub.com and Mobile experience Copilot in the cloud [37]; paid plans received that experience from 2026-10-28 [38].

How much does the Copilot cloud agent cost per month?

Individuals pay $10 (Pro), $39 (Pro+) or $100 (Max) each month; organizations pay $19 per Business user or $39 per Enterprise user; the Free tier has no cloud agent [7]. On top of the seat, usage is metered in AI Credits at $0.01 each [8].

Can the Copilot cloud agent use my own Anthropic or OpenAI key?

No. Bring-your-own-key is documented for the IDEs, Copilot CLI, the Copilot app and the SDK, not for the cloud agent [6]. Jiro runs on your Claude or ChatGPT login or API key, or a Cursor key [1].

Does Jiro have a 59-minute session limit like Copilot's Actions runs?

Copilot's Actions-based sessions stop at 59 minutes [9]. Jiro counts capacity in runtimes, which sleep while idle and wake for work, rather than by a per-run clock [41].

Which Copilot alternative works with GitLab or Bitbucket?

Cursor cloud agents support GitHub, GitLab, Bitbucket Cloud and Azure DevOps [53], Devin supports GitHub, GitLab, Bitbucket and Azure DevOps [57], and OpenHands Cloud supports GitHub, GitLab and Bitbucket [61]. Jiro itself needs GitHub, via its App or a PAT [46].

Can I keep copilot-instructions.md if I move to Jiro?

In effect, yes. A Nori skillset is translated into copilot-instructions.md as well as CLAUDE.md and AGENTS.md [76], so once its content is moved into a skillset both agents read the same rules.

Try Jiro

Start the 30-day free trial

No card needed. Or install Jiro's engineering skills into your own agent: npx nori-skillsets install public/jiro

Sources

  1. 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. https://docs.noriagentic.com/sessions/reference/supported-agents-and-compute (checked 2026-11-06)
  2. 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. https://noriagentic.com/#pricing (checked 2026-11-06)
  3. Nori's comparison table lists any model on your own keys with flat pricing and zero token markup. 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. 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. https://noriagentic.com/#faq (checked 2026-11-06)
  6. 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. https://docs.github.com/en/copilot/reference/ai-models/supported-models (checked 2026-11-06)
  7. 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. https://docs.github.com/en/copilot/get-started/plans (checked 2026-11-06)
  8. 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. https://docs.github.com/en/copilot/concepts/billing-and-usage/individuals/billing (checked 2026-11-06)
  9. 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. https://docs.github.com/en/copilot/concepts/agents/cloud-agent/about-cloud-agent (checked 2026-11-06)
  10. Copilot cloud agent opens draft pull requests on its own branch and cannot approve or merge them. https://docs.github.com/en/copilot/concepts/agents/cloud-agent/risks-and-mitigations (checked 2026-11-06)
  11. 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. https://learn.chatgpt.com/docs/pricing.md (checked 2026-11-06)
  12. 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. https://learn.chatgpt.com/docs/pricing.md (checked 2026-11-06)
  13. Codex cloud runs each task in a cloud VM, so work continues while your computer is asleep. https://learn.chatgpt.com/docs/environments/cloud-environments.md (checked 2026-11-06)
  14. In Codex cloud you review the changes, then commit or open a pull request when you are ready. https://learn.chatgpt.com/docs/cloud.md (checked 2026-11-06)
  15. Cloud sessions run Claude models and require an Anthropic account; they are not available through Bedrock, Google Cloud or other third-party providers. https://code.claude.com/docs/en/claude-code-on-the-web (checked 2026-11-06)
  16. 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. https://claude.com/pricing (checked 2026-11-06)
  17. 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. https://code.claude.com/docs/en/cloud-environments (checked 2026-11-06)
  18. 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. https://code.claude.com/docs/en/claude-code-on-the-web (checked 2026-11-06)
  19. Cloud agents use your plan's usage like other agent work, at the selected model's rate. https://cursor.com/docs/cloud-agent (checked 2026-11-06)
  20. 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. https://cursor.com/pricing (checked 2026-11-06)
  21. Cursor cloud agents run in isolated cloud VMs with full development environments, or move tool execution to self-hosted machines you manage. https://cursor.com/docs/cloud-agent (checked 2026-11-06)
  22. Cursor cloud agents produce merge-ready PRs with screenshots, videos and logs, and open a PR in each repo they touch. https://cursor.com/docs/cloud-agent (checked 2026-11-06)
  23. 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. https://docs.devin.ai/cli/models.md (checked 2026-11-06)
  24. 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. https://docs.devin.ai/admin/billing/self-serve.md (checked 2026-11-06)
  25. Devin Cloud runs sessions on Cognition-hosted machines; Devin CLI and Devin Desktop run locally. https://docs.devin.ai/llms.txt (checked 2026-11-06)
  26. Devin opens pull requests and automatically responds to comments on PRs its sessions are working on until the session is archived. https://docs.devin.ai/integrations/gh.md (checked 2026-11-06)
  27. OpenHands runs its own agent or Claude Code, Codex, Gemini, Pi, OpenCode or any ACP agent, and works with any LLM. https://github.com/OpenHands/OpenHands (checked 2026-11-06)
  28. Agent Canvas is free to self-host for a single user; Teams costs $500 a month for up to 10 users with extra seats at $50 a month each and a cap of 5,000 sessions a month, and Enterprise is custom. https://openhands.dev/pricing (checked 2026-11-06)
  29. OpenHands runs locally by default, with agents running locally, in Docker, on VMs, or optionally on OpenHands Cloud or OpenHands Enterprise infrastructure. https://github.com/OpenHands/OpenHands (checked 2026-11-06)
  30. Labelling a GitHub issue openhands or commenting @openhands starts a run that opens a PR if it resolves the issue. https://docs.openhands.dev/openhands/usage/cloud/github-installation (checked 2026-11-06)
  31. Copilot Pro includes 1,500 GitHub AI Credits a month, Pro+ 7,000 and Max 20,000; Business includes 1,900 credits per user a month and Enterprise 3,900 per user. https://docs.github.com/en/copilot/get-started/plans (checked 2026-11-06)
  32. Included AI Credits reset at 00:00 UTC on the first day of each calendar month and do not carry over; when they run out an individual can upgrade, set a dollar budget for additional usage or wait, while Business and Enterprise users must ask an administrator for more budget. https://docs.github.com/en/copilot/concepts/billing-and-usage/individuals/billing (checked 2026-11-06)
  33. 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. https://support.claude.com/en/articles/11049741-what-is-the-max-plan (checked 2026-11-06)
  34. Usage past a Pro, Max or Ultra plan's limits can continue as usage credits billed at standard API rates. https://support.claude.com/en/articles/12429409-extra-usage-for-paid-claude-plans (checked 2026-11-06)
  35. 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. https://cursor.com/docs/account/pricing/legacy-cursor-plans (checked 2026-11-06)
  36. Free and Pro allow up to 10 concurrent Devin sessions; Max and Teams have no concurrency cap. https://devin.ai/pricing (checked 2026-11-06)
  37. GitHub's Copilot cloud agent, now presented as Copilot in the cloud, researches a repository, plans changes and implements them in a GitHub cloud sandbox, a GitHub Actions environment or a remote environment you connect. https://docs.github.com/en/copilot/concepts/agents/cloud-agent/about-cloud-agent (checked 2026-11-06)
  38. Since 2026-10-28 paid Copilot plans get Copilot in the cloud on GitHub.com and GitHub Mobile, where the cloud sandbox can span several repositories in one organization and be resumed across clients, while the Actions-based environment is limited to a single repository and the GitHub website. https://docs.github.com/en/copilot/concepts/agents/cloud-agent/about-cloud-agent (checked 2026-11-06)
  39. Dependencies for the cloud agent are preinstalled by a .github/workflows/copilot-setup-steps.yml workflow whose single job must be named copilot-setup-steps, with timeout-minutes of at most 59, on Ubuntu x64 or Windows GitHub-hosted, larger or self-hosted runners. https://docs.github.com/en/copilot/how-tos/use-copilot-agents/cloud-agent/customize-the-agent-environment (checked 2026-11-06)
  40. Nori's default first-class Sessions agents are Claude Code, Codex, Cursor Agent, OpenCode and pi-acp. https://docs.noriagentic.com/sessions/reference/supported-agents-and-compute (checked 2026-11-06)
  41. Nori sizes plans by runtimes, which sleep when idle and wake on demand; the trial runs five for 30 days. https://noriagentic.com/#faq (checked 2026-11-06)
  42. A Nori organization can define a setup script that installs dependencies and prepares every session's environment. https://docs.noriagentic.com/start/what-can-nori-do#give-agents-a-real-working-environment (checked 2026-11-06)
  43. Nori hosts Jiro in its own cloud, and a dedicated BYOC deployment runs private EC2 session machines inside customer-controlled AWS infrastructure. https://docs.noriagentic.com/sessions/reference/supported-agents-and-compute (checked 2026-11-06)
  44. Each hosted Nori organization gets a single-tenant broker on its own subdomain. https://docs.noriagentic.com/sessions/concepts/security-and-credentials#organization-isolation (checked 2026-11-06)
  45. Nori logs every session and every action, and makes them searchable. https://noriagentic.com/for-security-leaders.html (checked 2026-11-06)
  46. 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. https://docs.noriagentic.com/sessions/guides/integrations#github (checked 2026-11-06)
  47. Codex cloud tasks start from ChatGPT on the web, mobile and desktop, or from the Codex CLI. https://learn.chatgpt.com/docs/cloud.md (checked 2026-11-06)
  48. Codex requests in Slack now start with @ChatGPT instead of @Codex. https://learn.chatgpt.com/docs/third-party/slack.md (checked 2026-11-06)
  49. Codex Cloud (Legacy) still powers Code Review and the GitHub and Linear integrations, and OpenAI plans to deprecate it. https://learn.chatgpt.com/docs/environments/cloud-environments.md (checked 2026-11-06)
  50. 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. https://code.claude.com/docs/en/claude-code-on-the-web (checked 2026-11-06)
  51. 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. https://code.claude.com/docs/en/slack (checked 2026-11-06)
  52. Cursor's help page on background agents says that in Cursor these are called Cloud Agents. https://cursor.com/help/ai-features/background-agents (checked 2026-11-06)
  53. 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. https://cursor.com/docs/cloud-agent (checked 2026-11-06)
  54. Cognition describes Devin as an autonomous software engineer that runs in the cloud or on your machine and tests its work in its own browser. https://devin.ai/ (checked 2026-11-06)
  55. Cognition introduced Devin Desktop as the next generation of Windsurf on 2026-06-02. https://cognition.com/blog/introducing-devin-desktop (checked 2026-11-06)
  56. Devin Cloud runs Cognition's Devin agent; Devin Desktop can also host other agents that speak ACP. https://cognition.com/blog/introducing-devin-desktop (checked 2026-11-06)
  57. Devin connects to GitHub, GitLab, Bitbucket and Azure DevOps repos and takes work from Slack, Microsoft Teams, Linear, Jira, PagerDuty and Pylon. https://docs.devin.ai/integrations/overview.md (checked 2026-11-06)
  58. Devin Automations start sessions from schedules, webhooks, and Slack, GitHub, GitLab, Linear, Jira, Pylon and PagerDuty events. https://docs.devin.ai/product-guides/automations.md (checked 2026-11-06)
  59. Devin Enterprise's Customer Dedicated Deployment runs the Devbox in a single-tenant VPC, while Devin's brain always stays in Cognition's cloud. https://docs.devin.ai/enterprise/deployment/overview.md (checked 2026-11-06)
  60. OpenHands is an MIT-licensed open-source, self-hosted control center for coding agents and automations. https://github.com/OpenHands/OpenHands (checked 2026-11-06)
  61. OpenHands Cloud works from its web UI and API, from GitHub, GitLab and Bitbucket repos, and from Slack. https://docs.openhands.dev/openhands/usage/cloud/cloud-ui (checked 2026-11-06)
  62. 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. https://code.claude.com/docs/en/costs (checked 2026-11-06)
  63. In Nori, one claimed machine serves one active session within the organization's fleet. https://docs.noriagentic.com/resources/faq#is-each-session-isolated (checked 2026-11-06)
  64. 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. https://docs.github.com/en/copilot/how-tos/use-copilot-agents/cloud-agent/create-a-pr (checked 2026-11-06)
  65. Copilot in the cloud sessions also start from automations that run on a schedule or on repository events, from GitHub Agentic Workflows written in Markdown and run as Actions workflows, from security campaigns that assign code scanning alerts, and from Agent Merge on failing workflows or merge conflicts. https://docs.github.com/en/copilot/concepts/agents/cloud-agent/about-cloud-agent (checked 2026-11-06)
  66. Nori work can start from Slack, the web dashboard, Discord, the Handroll terminal, schedules and webhooks, all reaching the same fleet. https://docs.noriagentic.com/start/what-can-nori-do#start-agent-work-from-the-tools-your-team-already-uses (checked 2026-11-06)
  67. 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. https://docs.noriagentic.com/sessions/guides/slack (checked 2026-11-06)
  68. Nori cron triggers run recurring jobs, reminders run once at a future time, and webhooks react to external events. https://docs.noriagentic.com/resources/faq#can-nori-run-tasks-without-a-person-online (checked 2026-11-06)
  69. Only users with write access can trigger the cloud agent; it pushes only to a copilot/ branch or the PR branch under existing branch protections, works behind a firewall, has its code checked by CodeQL, dependency review and secret scanning, needs a write-access user to approve Actions workflows on its PRs, and signs commits that link to session logs and audit log events. https://docs.github.com/en/copilot/concepts/agents/cloud-agent/risks-and-mitigations (checked 2026-11-06)
  70. Organization owners set Copilot policies under Settings > Copilot, including MCP servers in Copilot, which models beyond the basic ones are available, and Copilot in GitHub.com on Business or Enterprise, and an enterprise owner can fix any policy as enabled or disabled everywhere or for selected organizations. https://docs.github.com/en/copilot/how-tos/administer-copilot/manage-for-organization/manage-policies (checked 2026-11-06)
  71. Nori's dashboard exposes a catalog of MCP servers for project management, observability, analytics and other tools, and each server's own permissions still apply. https://docs.noriagentic.com/sessions/guides/integrations#managed-mcp-servers (checked 2026-11-06)
  72. On Pro, Pro+, Max, Business and Enterprise, Claude and Codex can be assigned GitHub tasks as third-party agents (public preview). https://docs.github.com/en/copilot/concepts/agents/about-third-party-agents (checked 2026-11-06)
  73. Nori handles more than 25% of Miravoice's merged pull requests, and Miravoice's CTO says it easily saves him five hours a week. https://noriagentic.com/newsletter/2026-09-18-miravoice-case-study.html (checked 2026-11-06)
  74. The Miravoice team set up an agent entirely in the cloud from their phones. https://noriagentic.com/#miravoice-story (checked 2026-11-06)
  75. Corvus's CTO saw a production bug at a client dinner, had Nori look into it and deploy a fix. https://noriagentic.com/newsletter/2026-05-19-corvus-case-study.html (checked 2026-11-06)
  76. A Nori skillset is written once and translated into each agent's own files: CLAUDE.md for Claude Code, GEMINI.md for Gemini CLI, copilot-instructions.md for GitHub Copilot and AGENTS.md for the rest. https://docs.noriagentic.com/skillsets/concepts/translation-and-agents (checked 2026-11-06)
  77. Nori is led by CEO Amol Kapoor, a former Google Research scientist, and CTO and co-founder Clifford Ressel. https://ai.engineer/speakers/amol-kapoor (checked 2026-11-06)
  78. Amol Kapoor's Google-affiliated research includes MixHop (ICML 2019), N-GCN and a 2020 spatio-temporal graph neural network for COVID-19 forecasting. https://scholar.google.com/citations?user=SC_-qB8AAAAJ (checked 2026-11-06)
  79. Clifford Ressel is Nori's CTO and co-founder. https://ai.engineer/orgs/nori (checked 2026-11-06)