--- title: "Cork vs Hoplite: cloud agents vs coordinating the agents you have · Cork" description: "Cork vs Hoplite: Hoplite runs coding agents in isolated cloud environments that open PRs. Cork keeps your local parallel agents from colliding." canonical: "https://corkcli.com/blog/cork-vs-hoplite" last_updated: "2026-11-14" published: "2026-10-03" author: "Kevin Vera" --- # Cork vs Hoplite: cloud agents vs coordinating the agents you have Cork vs Hoplite: Hoplite runs coding agents in isolated cloud environments that open PRs. Cork keeps your local parallel agents from colliding. By Kevin Vera · October 3, 2026 · [Comparisons](https://corkcli.com/blog/c/comparisons) · 7 min read How is Cork different from Hoplite? Hoplite runs coding agents for you in the cloud: each task gets an isolated development environment where the agent edits code, runs tests, drives a browser to verify the change and opens a pull request. Cork runs nothing; it coordinates the agents you already run (Claude Code, Codex, Cursor) on your own machine, so that several of them, and the people running them, do not overwrite each other on one codebase. Put simply, Hoplite is the agent and the machine, while Cork is the layer that keeps parallel agents and teammates aware of each other's code. Whether you need both depends on where your agents run. ## Cork vs Hoplite at a glance | Topic | Cork | Hoplite (as of October 2026) | |---|---|---| | What it is | A [coordination layer](https://corkcli.com/blog/coordination-layer-ai-coding-agents) for multiple AI coding agents on one codebase | A cloud coding agent service that verifies changes and opens PRs | | Where agents run | Where they already run, on your machine | Isolated cloud development environments, one per thread | | Agents covered | Claude Code, Codex, Cursor (hooks), plus any MCP client | Hoplite's own in-house agent harness, with a choice of models | | What is shared | Live code changes, signatures, callers, notes anchored to code | Task threads, verification evidence and pull requests | | Conflict handling | Write gate, live awareness, [semantic lint](https://corkcli.com/blog/semantic-lint), rebase-test-push landing | Each task works in its own environment; overlap is resolved at PR review | | Install | `curl -fsSL https://corkcli.com/install.sh \| sh` | Hosted sign-up with a GitHub App connection | | Open source | Cork Free runs locally with no account | We found no open source release | | Pricing | Free (local, and 10 agent hours a month in the cloud); $20 a month per 100 agent hours, pooled | Free tier (card required); Pro $99/seat/mo; Enterprise custom | ## What is Hoplite? Hoplite (a Y Combinator S26 company) describes its product as "cloud coding agents that work in isolated development environments, verify their changes, and open pull requests." Per its website and the founders' Launch HN post: - **Cloud environments per task.** Each thread gets an environment where the agent can inspect code, edit files, run tests and start the app. - **Verification.** Agents get a persistent Chromium session to check their work visually, and teams can save flows in a QA library to replay them later without rerunning an agent. - **Your local setup, ported.** Onboarding brings over local sessions, memories and MCP servers, so cloud work starts with the same context. - **Its own harness.** The founders say they considered using Codex or Claude Code but built the agent harness in house; you can also connect a Codex subscription to use OpenAI models. - **Scale.** The founders' stated vision is reviewing product output while running hundreds of agents concurrently. It connects to repositories through a GitHub App and supports MCP for external tools, with approval for sensitive actions. ## What is Cork? Cork (corkcli.com, by Lithometric) keeps several AI coding agents, and the people running them, from overwriting each other on one codebase, and hands each agent what the others already learned. It does not run agents or host machines; your agents keep running where they already run, and Cork coordinates them through hooks in Claude Code, Codex and Cursor (no extra model calls) plus an MCP server any MCP client can use. - **Briefing** with the exact signatures, callers and recent changes of the code an agent is about to touch. - **Live awareness** of renames and signature changes on the agent's next step. - **[Write gate](https://corkcli.com/blog/write-gate)** that holds an edit against code another agent is changing until your agent has their version. - **Semantic lint** from teammates' renames and signature changes. - **Landing**: `git push` rebases onto teammates' commits, reruns the tests and pushes, in one step. - **Agent messaging**, plus `get_symbol`, `blast_radius`, `recap` and `remember`/`recall`. Cork does not provide cloud VMs, sandboxes, a code editor or app previews. ## Where Hoplite stands out - **No local machine needed.** Close your laptop and the agent keeps working. Cork does not host anything. - **Built-in verification.** Browser-driven checks and a replayable QA library go beyond "tests passed." - **PRs as the output.** Work arrives as a reviewed pull request, ready for your normal review flow. - **Scale in the cloud.** Many concurrent sessions without tying up your hardware. - **Smooth onboarding** that carries your local sessions, memories and MCP servers into the cloud. ## Where Cork wins - **Your agents, your machine.** Keep using Claude Code, Codex or Cursor with your own setup; nothing moves to a hosted environment. - **Collisions caught while the work happens.** Isolated environments avoid stepping on each other mid-task, but parallel changes still meet at merge time as agent merge conflicts. Cork warns agents live and gates conflicting edits before they become a PR fight. - **Teams of people, not just agents.** On the paid plan, agents across teammates' machines see each other's live changes. - **Price.** Cork bills pooled agent hours, $20 a month per 100 hours across the workspace, not seats, against Hoplite Pro's $99 per seat (which includes model usage, a different thing to buy); the two are not per-person equivalents. - **Measured savings.** In Cork's benchmark (12 Claude Code agents at once on one repository, hard tasks), agents with Cork used 53% fewer tokens than the same agents without it. ## Can you use both? Partly. Cork's hooks live inside Claude Code, Codex and Cursor; Hoplite runs its own harness in its own cloud, so Cork does not coordinate Hoplite's agents directly. The realistic combination is a team that hands backlog tasks to Hoplite while developers keep running local agents with Cork: Cork keeps the local agents and teammates from colliding, and Cork's landing step rebases onto whatever Hoplite PRs have merged and reruns the tests before pushing. If all of your agent work moves to Hoplite, Cork adds little. ## Pricing compared **Cork:** Cork Free covers unlimited agents and repos on one machine, no account, with all data on the machine. Cloud Free (with an account) is 10 agent hours a month per account (30 bonus hours the first month), unlimited repos and agents, 30-day history. Beyond that you pay for agent hours, not seats: $20 a month per unit of 100 hours, pooled across the workspace, graduated to $17 (units 11 to 50) and $15 (51 and up), with 15% off a year. Every member is a player, with teammates' machines, prompts matched to code by meaning, shared knowledge and 90-day history. Enterprise starts from about 100 units. **Hoplite (as of October 2026):** Free is $0 for one person with 5 concurrent cloud sessions (2 CPU / 8 GB), 2 projects and a card required (nothing charged). Pro is $99 per seat per month ($82.50 on yearly billing) with a monthly usage allowance, team seats and higher limits. Enterprise is custom. Hoplite's price includes cloud compute and model usage; Cork's does not, because your agents run on your own subscriptions. ## Bottom line Hoplite is a strong choice if you want agents that run in the cloud, verify their own work and hand you a pull request. Cork is for teams who keep their agents local and need [parallel AI coding agents](https://corkcli.com/blog/parallel-ai-coding-agents) and teammates to stop overwriting each other. They overlap less than they seem: one is a cloud agent, the other is coordination for the agents you already have. ## Frequently asked questions ### Is Cork an alternative to Hoplite? Only if your question is "how do I run more agents in parallel without chaos." Hoplite answers it by moving agents to isolated cloud environments; Cork answers it by coordinating the local agents you already use. ### Does Hoplite stop agents from overwriting each other? Each Hoplite task runs in its own isolated environment and ends in a pull request, so agents do not overwrite each other mid-task. Overlapping changes are resolved at review and merge time. ### Can Hoplite run Claude Code? Not as its core agent. The founders say they built their own harness instead of using Codex or Claude Code, though Hoplite offers Anthropic and OpenAI models and lets you connect a Codex subscription. ### Is Hoplite open source? We found no open source release as of October 2026. ### Does Cork provide cloud sandboxes like Hoplite? No. Cork does not provide cloud VMs, sandboxes or app previews; it coordinates agents wherever they already run. ## Sources Checked October 2026: - - - - ## Related reading - [Cork vs Cursor cloud agents: run agents in VMs, or keep them from colliding](https://corkcli.com/blog/cork-vs-cursor-cloud-agents): Cork vs Cursor cloud agents: Cursor runs agents in isolated cloud VMs that open PRs; Cork keeps parallel agents from overwriting shared code. - [Cork vs Sculptor: an app for parallel agents, or a layer that keeps them from colliding](https://corkcli.com/blog/cork-vs-sculptor): Cork vs Sculptor: Imbue's Sculptor runs coding agents in parallel in isolated workspaces. Cork coordinates parallel agents and teammates on shared code. - [Cork vs Emdash: an agent workbench vs a coordination layer](https://corkcli.com/blog/cork-vs-emdash): Cork vs Emdash: Emdash is an open-source app that runs coding agents in parallel worktrees; Cork keeps those agents from colliding on shared code. - [Cork vs Warp: orchestrating agents in the cloud, or coordinating them on one codebase](https://corkcli.com/blog/cork-vs-warp): Cork vs Warp: Warp's Oz runs and orchestrates cloud coding agents at scale. Cork keeps parallel AI coding agents from overwriting each other's code. [All articles](https://corkcli.com/blog.md) · [Cork in Markdown](https://corkcli.com/index.md) · [llms.txt](https://corkcli.com/llms.txt) This page in HTML: https://corkcli.com/blog/cork-vs-hoplite