# Rush > Rush runs agent work unattended and hands back the evidence. Send a request with an API key; an agent does the work in its own cloud sandbox, with a computer and a browser, and returns a pull request, a report, files or an answer, beside the full record of how it got there. Rush is made by Phoenix (https://phoenix.so), which builds secure, scalable infrastructure for autonomous work. The API is at https://api.phoenix.so; the MCP server at https://api.phoenix.so/mcp hands the same API to an agent in any MCP client, and the console at https://cloud.getrush.ai hands it to a person. ## Get started - Agent skill: https://getrush.ai/SKILL.md teaches an agent to hand work to Rush (a key, a session, its events, a follow-up, the result, a webhook). For Claude Code: `mkdir -p ~/.claude/skills/rush && curl -fsSL https://getrush.ai/SKILL.md -o ~/.claude/skills/rush/SKILL.md` - Get an API key: an org admin creates one on the console's Developers page (https://cloud.getrush.ai/console/developers), which also shows the org's path. Send it as `Authorization: Bearer $RUSH_API_KEY`. - [API quickstart](https://getrush.ai/docs/quickstart.md): get a key, start a session, stream its events, follow up, read the result, get a webhook - [Rush MCP server](https://getrush.ai/docs/concepts/mcp-server.md): Hand Rush a task from claude.ai, ChatGPT, Claude Code, Cursor or Codex: connect api.phoenix.so/mcp by signing in or with an API key, then start runs as tool calls. - OpenAPI document: https://getrush.ai/openapi.json (API catalog: https://getrush.ai/.well-known/api-catalog) Start a session: ```bash curl -X POST "https://api.phoenix.so/o/$RUSH_ORG/sessions" \ -H "Authorization: Bearer $RUSH_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "agent": "claude", "input": "Write a Python function that parses ISO 8601 durations like P1DT2H30M, with tests. Run the tests and report the result." }' ``` ## How a run works 1. Send one request with an API key, from an MCP client or from the console: the task in plain words, plus a repository when the work touches code. 2. The run starts in its own cloud computer with the project's context (and the repository, when there is one), shared with no other run. 3. Follow the work live, send a follow-up while it runs, or cancel it. A webhook tells your server when a turn ends. 4. The run comes back with the result and every step the agent took: for code, a branch and a pull request with checks; otherwise an answer, a report or files. A run parks with its memory and files intact when a turn ends, costs nothing while parked, and resumes when you write back. The agent is Claude Code, OpenAI Codex, or OpenCode, picked per run with the `agent` field. ## Pricing - Free: $0. 5 agent-hours a month. Up to 2 of them in any one week. One agent at a time. Cloud computer only: connect your own model account. Your first $1 of models on us. No card required. - Pro: $20 a month. 40 agent-hours a month. Up to 15 of them in any one week. 4 agents at once. Cloud computer only: connect your own model account. Cancel any time. - Team: $200 a month. 400 agent-hours a month. Up to 150 of them in any one week. 12 agents at once. Cloud computer only: connect your own model account. Cancel any time. You pay for the cloud the agent runs in, measured in agent-hours, not for seats. Bring your Claude Code or Codex seat or your own API key; the run still draws on the plan's agent-hours. ## Docs The home page (at `/index.md`), pricing, every docs, comparison, guide, lesson, task and blog page, and the compare, guides, learn and blog indexes have a Markdown copy at the same path with `.md` appended; the links below go to them. A page also answers in Markdown to a request whose `Accept` header prefers `text/markdown`. - [Docs index](https://getrush.ai/docs.md): what Rush does and a link to every docs page - [Runs and sessions](https://getrush.ai/docs/concepts/runs-and-sessions.md): A session is one agent run. It moves through a fixed set of statuses, takes follow-ups between turns and keeps its sandbox while it is parked. - [Your first run from code](https://getrush.ai/docs/learn/first-api-run.md): Create an API key, start a session with one request, follow it live with curl, read its answer and send a follow-up into the same sandbox. - [Your first run in the console](https://getrush.ai/docs/start/console.md): Sign in at cloud.getrush.ai, connect GitHub, create a Project and start a run, then watch it work and open the pull request it made. - [Act on the result with a webhook](https://getrush.ai/docs/learn/act-on-the-result.md): Point your key at an endpoint, verify each signed POST, and handle every finished turn exactly once, even when a delivery is retried. - [Projects and secrets](https://getrush.ai/docs/concepts/projects.md): A Project ties a repository, its secrets and its shared connections together inside your org. Sessions in it get all three without naming them. - [Give a run files, skills and connections](https://getrush.ai/docs/learn/files-skills-connections.md): Start a run with files in its sandbox, skills loaded from a URL, the tools it may call and the repository or issue it works from. - [Results and proof](https://getrush.ai/docs/concepts/results-and-proof.md): Every run ends with a record you can check: the answer, the branch and pull request, the files it saved, its transcript and what it used. - [Run a saved agent from code](https://getrush.ai/docs/learn/saved-agents-from-code.md): Save a harness with its model, instructions, skills and connections once, then start each run with its id and a task. - [Watch and steer live](https://getrush.ai/docs/concepts/watch-and-steer.md): Follow a run as it works, send it the next instruction, stop it, or take over its desktop. Every surface reads the same event stream. - [Isolation, plainly](https://getrush.ai/docs/concepts/isolation.md): Each run gets its own microVM, each org sees only its own resources, and secrets reach a run only in the form you chose. What that means in practice. - [Work that starts itself](https://getrush.ai/docs/learn/work-that-starts-itself.md): Turn a prompt into a Project automation that runs on a GitHub event, a Linear issue, a schedule or a webhook, and read what each run did. - [Models and your access](https://getrush.ai/docs/concepts/models.md): API, MCP and template-agent runs use Rush's models, billed per token; console runs use your own account. Pick a model and effort from the list your org can run. - [Steer a run as a team](https://getrush.ai/docs/learn/steer-a-run-as-a-team.md): Invite teammates to your org, see every run in a Project with who started it, queue follow-ups mid-turn and take control of a run's browser. - [Automations](https://getrush.ai/docs/concepts/automations.md): An automation is a saved prompt in a Project with triggers that say when it runs: a GitHub event, a schedule, a webhook call or a Linear issue change. - [Saved agents](https://getrush.ai/docs/concepts/saved-agents.md): Save a harness with its model, effort, instructions, skills and connections once, then start every run of it by sending its id as agent. - [Connections](https://getrush.ai/docs/concepts/connections.md): Sign in to a tool or MCP server once, then name it on any run. Skills and resources come in by URL the same way. - [Webhooks](https://getrush.ai/docs/concepts/webhooks.md): Give an API key an endpoint and every run it starts posts a signed event there when a turn ends, so you never hold a stream or poll. - [Usage and billing](https://getrush.ai/docs/concepts/usage-and-billing.md): Rush meters compute while a run is awake and tokens on Rush's models. Every session records its tokens and cost; limits answer with a typed code. - [Community agents](https://getrush.ai/docs/concepts/community-agents.md): Companies on Rush share saved agents they run. Add one and it becomes your own agent, running on your own apps, while its instructions stay with its maker. - [Use case: Review every pull request](https://getrush.ai/docs/use-cases/review-pull-requests.md): A review on the PR. Each pull request gets an agent that checks it out in its own sandbox, runs the tests and posts a review with the evidence. - [Use case: Fix a bug from an alert](https://getrush.ai/docs/use-cases/fix-bugs-from-alerts.md): A PR with checks. Your alerting posts to a webhook trigger; an agent reproduces the failure, fixes it and opens a pull request with the checks passing. - [Use case: Run an agent per customer task](https://getrush.ai/docs/use-cases/agent-per-customer-task.md): A typed answer, by webhook. Your backend starts one session per customer request with an API key, and Rush posts the result to your webhook when the turn ends. - [Use case: Produce a report or file, no repo](https://getrush.ai/docs/use-cases/reports-without-a-repo.md): A file you download. Send a request with no repository; the agent works in its own computer and returns a file as one of the session's artifacts. - [Use case: Write the missing tests](https://getrush.ai/docs/use-cases/write-missing-tests.md): A PR with passing checks. A nightly schedule starts an agent that finds untested code, writes the tests and opens a pull request by morning. - [Use case: Investigate an incident](https://getrush.ai/docs/use-cases/investigate-incidents.md): A report with evidence. Hand the agent an incident and read access to its sources; it comes back with a report that cites what it found, and no code change. - [Use case: Turn a Linear label into a PR](https://getrush.ai/docs/use-cases/linear-label-to-pr.md): A PR linked to the issue. Label a Linear issue and an agent picks it up, does the work in the Project's repository and opens a pull request. - [Use case: Upgrade dependencies weekly](https://getrush.ai/docs/use-cases/weekly-dependency-upgrades.md): A PR, or needs_review. Once a week an agent bumps your dependencies, runs the tests and opens a pull request, or stops at needs_review and says why. - [Use case: Draft outreach from your CRM](https://getrush.ai/docs/use-cases/draft-outreach-from-your-crm.md): A sourced draft, approval in Slack. For each prospect, an agent researches the company, saves a sourced first message in your CRM and asks you in Slack to approve it. A person sends it. - [Use case: Steer a run as a team](https://getrush.ai/docs/use-cases/steer-a-run-as-a-team.md): One run, many hands. A teammate starts a run; everyone in the org can read it as it works, and its starter or an org admin steers it with follow-ups or by taking control of its browser. - [Error codes](https://getrush.ai/docs/api/errors.md): every code, its status and the operations that send it - [API reference](https://getrush.ai/docs/api.md): every public operation - [List sessions](https://getrush.ai/docs/api/list-sessions.md): `GET /o/{org}/sessions` - [Start a session](https://getrush.ai/docs/api/create-session.md): `POST /o/{org}/sessions` - [Get one session](https://getrush.ai/docs/api/get-session.md): `GET /o/{org}/sessions/{session}` - [Cancel a session](https://getrush.ai/docs/api/cancel-session.md): `POST /o/{org}/sessions/{session}/cancel` - [Start a session (console, by Project)](https://getrush.ai/docs/api/create-project-session.md): `POST /o/{org}/p/{project}/sessions` - [List the turns of a session](https://getrush.ai/docs/api/list-turns.md): `GET /o/{org}/p/{project}/sessions/{session}/turns` - [Follow a session live](https://getrush.ai/docs/api/stream-session.md): `GET /o/{org}/sessions/{session}/events` - [Read a session's transcript](https://getrush.ai/docs/api/get-session-messages.md): `GET /o/{org}/sessions/{session}/messages` - [Send a follow-up](https://getrush.ai/docs/api/post-session-message.md): `POST /o/{org}/sessions/{session}/messages` - [Download one artifact](https://getrush.ai/docs/api/get-session-artifact.md): `GET /o/{org}/sessions/{session}/artifacts/{name}` - [List the files the session's pull request changed](https://getrush.ai/docs/api/get-session-changes.md): `GET /o/{org}/p/{project}/sessions/{session}/changes` - [List the files a session produced](https://getrush.ai/docs/api/list-session-artifacts.md): `GET /o/{org}/p/{project}/sessions/{session}/artifacts` - [List the org's Project secrets, without values.](https://getrush.ai/docs/api/list-secrets.md): `GET /o/{org}/secrets` - [Save a secret. The value is sealed and never returned.](https://getrush.ai/docs/api/create-secret.md): `POST /o/{org}/secrets` - [Replace a secret's value or note, or a sign-in's site.](https://getrush.ai/docs/api/update-secret.md): `PATCH /o/{org}/secrets/{secret}` - [Delete a secret. A session that names it afterwards fails instead of starting without it.](https://getrush.ai/docs/api/delete-secret.md): `DELETE /o/{org}/secrets/{secret}` - [List the org's API keys](https://getrush.ai/docs/api/list-api-keys.md): `GET /o/{org}/api-keys` - [Create an API key](https://getrush.ai/docs/api/create-api-key.md): `POST /o/{org}/api-keys` - [Set, change or remove the key's webhook](https://getrush.ai/docs/api/update-api-key-webhook.md): `PATCH /o/{org}/api-keys/{key}` - [Revoke an API key](https://getrush.ai/docs/api/revoke-api-key.md): `DELETE /o/{org}/api-keys/{key}` - [Rotate an API key's secret](https://getrush.ai/docs/api/rotate-api-key.md): `POST /o/{org}/api-keys/{key}/rotate` - [List the harnesses, models and connectors the org can run](https://getrush.ai/docs/api/get-agents-available.md): `GET /o/{org}/agents/available` - [A run started with an API key finished a turn](https://getrush.ai/docs/api/run-finished.md): webhook - [List the org's Projects](https://getrush.ai/docs/api/list-projects.md): `GET /o/{org}/p` - [Get a Project by handle or id](https://getrush.ai/docs/api/get-project.md): `GET /o/{org}/p/{project}` - [List the GitHub App installations the caller can use and their repositories.](https://getrush.ai/docs/api/get-github-connection.md): `GET /o/{org}/github` - [List saved agents you can use, newest first.](https://getrush.ai/docs/api/list-saved-agents.md): `GET /o/{org}/agents` - [Save an agent.](https://getrush.ai/docs/api/create-saved-agent.md): `POST /o/{org}/agents` - [Get a saved agent you can use.](https://getrush.ai/docs/api/get-saved-agent.md): `GET /o/{org}/agents/{agent}` - [Edit a saved agent.](https://getrush.ai/docs/api/update-saved-agent.md): `PATCH /o/{org}/agents/{agent}` - [Delete a saved agent.](https://getrush.ai/docs/api/delete-saved-agent.md): `DELETE /o/{org}/agents/{agent}` - [List the accounts and MCP servers connected to one agent made from a template.](https://getrush.ai/docs/api/list-agent-connections.md): `GET /o/{org}/agents/{agent}/connections` - [Start a sign-in that connects an account or MCP server to one agent.](https://getrush.ai/docs/api/create-agent-connection.md): `POST /o/{org}/agents/{agent}/connections` - [Read one of an agent's connections, or the sign-in that makes it.](https://getrush.ai/docs/api/get-agent-connection.md): `GET /o/{org}/agents/{agent}/connections/{connection}` - [Disconnect one of an agent's connections, or cancel its sign-in.](https://getrush.ai/docs/api/delete-agent-connection.md): `DELETE /o/{org}/agents/{agent}/connections/{connection}` - [Read one agent listed in Community](https://getrush.ai/docs/api/get-community-agent.md): `GET /o/{org}/agents/community/{agent}` - [List the agents companies listed in Community](https://getrush.ai/docs/api/list-community-agents.md): `GET /o/{org}/agents/community` - [Download one file the session was given](https://getrush.ai/docs/api/get-session-attachment.md): `GET /o/{org}/sessions/{session}/attachments/{id}` - [Get an upload URL for each file of a session's input.](https://getrush.ai/docs/api/create-files.md): `POST /o/{org}/files` ## Key pages - [Home](https://getrush.ai/index.md): what Rush is, a real run, how it works, pricing and FAQ - [How a run works](https://getrush.ai/cloud/): the run lifecycle, isolation and credentials in detail - [Pricing](https://getrush.ai/pricing.md): Free, Pro and Team plans - [Compare](https://getrush.ai/compare.md): Rush next to Devin, Cursor Cloud Agents, GitHub Copilot, Claude Code on the web and Codex - [Status](https://getrush.ai/status/): component health and incident history - [What's new](https://getrush.ai/changelog/): what changed in the API, the MCP server and the console, one week at a time ## Comparisons - [Devin alternatives (2026)](https://getrush.ai/compare/devin-alternatives.md): Tools that take a ticket into an isolated cloud box and return a pull request: Rush, Cursor Cloud Agents, Copilot, Codex, Jules and more, sourced September 2026. - [Cursor Cloud Agents alternatives (2026)](https://getrush.ai/compare/cursor-cloud-agents-alternatives.md): Alternatives to Cursor Cloud Agents, formerly Background Agents: Rush, Claude Code on the web, Codex cloud, Copilot, Devin and more, sourced September 2026. - [GitHub Copilot cloud agent alternatives (2026)](https://getrush.ai/compare/github-copilot-coding-agent-alternatives.md): Alternatives to GitHub Copilot cloud agent, once Copilot coding agent: Rush, Claude Code on the web, Cursor Cloud Agents, Devin, Jules and Codex, September 2026. - [Claude Code on the web alternatives (2026)](https://getrush.ai/compare/claude-code-web-alternatives.md): Alternatives to Claude Code on the web: Rush, Cursor Cloud Agents, Codex cloud, Copilot cloud agent and Devin, each from the vendor's pages, checked September 2026. - [Rush vs Devin (2026)](https://getrush.ai/compare/rush-vs-devin.md): Rush runs Claude Code, Codex or OpenCode in an isolated Project. Devin is Cognition's own agent. Same job, a different engine and meter. Sources from September 2026. - [Rush vs Cursor Cloud Agents (2026)](https://getrush.ai/compare/rush-vs-cursor-cloud-agents.md): Rush runs Claude Code, Codex, or OpenCode. Cursor Cloud Agents run Cursor's Agent in Firecracker VMs. Sources retrieved 15 September 2026. - [Rush vs GitHub Copilot cloud agent (2026)](https://getrush.ai/compare/rush-vs-github-copilot.md): Rush vs GitHub Copilot cloud agent. Same GitHub PR, different engine, different session cap. Sources retrieved 15 September 2026. - [Rush vs Claude Code on the web (2026)](https://getrush.ai/compare/rush-vs-claude-code-web.md): Same Claude Code harness. Anthropic hosts one VM. Rush hosts a Project that can also run Codex and OpenCode. Sources retrieved 15 September 2026. - [Devin vs Cursor Cloud Agents (2026)](https://getrush.ai/compare/devin-vs-cursor.md): Cognition's Devin versus Cursor Cloud Agents. Two vendor agents, two meters, one job: isolated cloud to pull request. Sources retrieved 15 September 2026. - [Devin review (2026)](https://getrush.ai/compare/devin-review.md): A sourced review of Cognition's Devin: what it is, what it costs, what independent tests said, and when to pick something else. Retrieved 15 September 2026. - [Cursor Cloud Agents review (2026)](https://getrush.ai/compare/cursor-cloud-agents-review.md): A sourced review of Cursor Cloud Agents (formerly Background Agents): pricing, isolation, MCP limits, and what Cursor's own posts claim. Retrieved 15 September 2026. - [Best cloud coding agents (2026)](https://getrush.ai/compare/best-cloud-coding-agents.md): A sourced map of cloud coding agents that open a pull request. Scenario routing, not a crown. Verified 15 September 2026. - [Best way to run Claude Code in the cloud (2026)](https://getrush.ai/compare/best-claude-code-cloud.md): Claude Code on the web, Rush, and GitHub's third-party Claude agent preview. Same harness, three runtimes. Retrieved 15 September 2026. - [Cloud coding agents with MCP (2026)](https://getrush.ai/compare/coding-agents-with-mcp.md): How MCP actually loads on Devin, Cursor Cloud Agents, Copilot cloud agent, Claude Code on the web, Factory, Jules, Amp, and Rush. Retrieved 15 September 2026. - [Cloud coding agents for engineering teams (2026)](https://getrush.ai/compare/engineering-teams.md): How a small startup engineering team runs Claude Code, Codex, or OpenCode in the cloud with a shared Project. Rush, with the vendor alternatives named. - [How to move off Devin (2026)](https://getrush.ai/compare/move-from-devin.md): What you keep when you leave Cognition's Devin: git remotes, the API export, the GitHub App, and privacy requests. Then how to run the next Task on Rush. - [Migrate from Cursor Cloud Agents to Rush](https://getrush.ai/compare/migrate-from-cursor-cloud-agents.md): How to move work from Cursor Cloud Agents to Rush: what to keep, what carries over and the first run. Steps from Anysphere's own docs, checked 15 September 2026. - [Migrate from GitHub Copilot cloud agent to Rush](https://getrush.ai/compare/migrate-from-github-copilot.md): How to move work from GitHub Copilot cloud agent to Rush: what to keep, what carries over and the first run. Steps from GitHub's own docs, checked 15 September 2026. - [Migrate from Claude Code on the web to Rush](https://getrush.ai/compare/migrate-from-claude-code-web.md): How to move work from Claude Code on the web to Rush: what to keep, what carries over and the first run. Steps from Anthropic's own docs, checked 15 September 2026. - [Rush vs Codex cloud (2026)](https://getrush.ai/compare/rush-vs-codex-cloud.md): Rush vs Codex cloud: what runs, isolation, pull requests, MCP and price, each cited from OpenAI's own pages, checked 15 September 2026. - [Codex cloud alternatives (2026)](https://getrush.ai/compare/codex-cloud-alternatives.md): Codex cloud alternatives: Rush, Claude Code on the web and 5 more. Price, isolation and pull requests from each vendor, checked 15 September 2026. - [Migrate from Codex cloud to Rush](https://getrush.ai/compare/migrate-from-codex-cloud.md): How to move work from Codex cloud to Rush: what to keep, what carries over and the first run. Steps from OpenAI's own docs, checked 15 September 2026. - [Rush vs Factory Droids (2026)](https://getrush.ai/compare/rush-vs-factory-droids.md): Rush vs Factory Droids: what runs, isolation, pull requests, MCP and price, each cited from Factory's own pages, checked 15 September 2026. - [Factory Droids alternatives (2026)](https://getrush.ai/compare/factory-droids-alternatives.md): Factory Droids alternatives: Rush, Devin, Cursor Cloud Agents and 4 more. Price, isolation and pull requests from each vendor, checked 15 September 2026. - [Migrate from Factory Droids to Rush](https://getrush.ai/compare/migrate-from-factory-droids.md): How to move work from Factory Droids to Rush: what to keep, what carries over and the first run. Steps from Factory's own docs, checked 15 September 2026. - [Rush vs Google Jules (2026)](https://getrush.ai/compare/rush-vs-jules.md): Rush vs Google Jules: what runs, isolation, pull requests, MCP and price, each cited from Google's own pages, checked 15 September 2026. - [Google Jules alternatives (2026)](https://getrush.ai/compare/jules-alternatives.md): Google Jules alternatives: Rush, Codex cloud and 5 more. Price, isolation and pull requests from each vendor, checked 15 September 2026. - [Migrate from Google Jules to Rush](https://getrush.ai/compare/migrate-from-jules.md): How to move work from Google Jules to Rush: what to keep, what carries over and the first run. Steps from Google's own docs, checked 15 September 2026. - [Rush vs Amp orbs (2026)](https://getrush.ai/compare/rush-vs-amp.md): Rush vs Amp orbs: what runs, isolation, pull requests, MCP and price, each cited from Amp Inc.'s own pages, checked 15 September 2026. - [Amp orbs alternatives (2026)](https://getrush.ai/compare/amp-alternatives.md): Amp orbs alternatives: Rush, Factory Droids, Cursor Cloud Agents and 4 more. Price, isolation and pull requests from each vendor, checked 15 September 2026. - [Migrate from Amp orbs to Rush](https://getrush.ai/compare/migrate-from-amp.md): How to move work from Amp orbs to Rush: what to keep, what carries over and the first run. Steps from Amp Inc.'s own docs, checked 15 September 2026. - [Rush vs E2B (2026)](https://getrush.ai/compare/rush-vs-e2b.md): Rush vs E2B: what runs, isolation, pull requests, MCP and price, each cited from E2B's own pages, checked 17 October 2026. - [E2B alternatives (2026)](https://getrush.ai/compare/e2b-alternatives.md): E2B alternatives: Rush, Daytona, Modal Sandboxes, Runloop and 3 more. Price, isolation and pull requests from each vendor, checked 17 October 2026. - [Migrate from E2B to Rush](https://getrush.ai/compare/migrate-from-e2b.md): How to move work from E2B to Rush: what to keep, what carries over and the first run. Steps from E2B's own docs, checked 17 October 2026. - [Rush vs Daytona (2026)](https://getrush.ai/compare/rush-vs-daytona.md): Rush vs Daytona: what runs, isolation, pull requests, MCP and price, each cited from Daytona's own pages, checked 17 October 2026. - [Daytona alternatives (2026)](https://getrush.ai/compare/daytona-alternatives.md): Daytona alternatives: Rush, E2B, Modal Sandboxes, Runloop and 3 more. Price, isolation and pull requests from each vendor, checked 17 October 2026. - [Migrate from Daytona to Rush](https://getrush.ai/compare/migrate-from-daytona.md): How to move work from Daytona to Rush: what to keep, what carries over and the first run. Steps from Daytona's own docs, checked 17 October 2026. - [Rush vs Modal Sandboxes (2026)](https://getrush.ai/compare/rush-vs-modal.md): Rush vs Modal Sandboxes: what runs, isolation, pull requests, MCP and price, each cited from Modal's own pages, checked 17 October 2026. - [Modal Sandboxes alternatives (2026)](https://getrush.ai/compare/modal-alternatives.md): Modal Sandboxes alternatives: Rush, E2B, Daytona, Runloop and 3 more. Price, isolation and pull requests from each vendor, checked 17 October 2026. - [Rush vs Runloop (2026)](https://getrush.ai/compare/rush-vs-runloop.md): Rush vs Runloop: what runs, isolation, pull requests, MCP and price, each cited from Runloop AI's own pages, checked 17 October 2026. - [Runloop alternatives (2026)](https://getrush.ai/compare/runloop-alternatives.md): Runloop alternatives: Rush, E2B, Daytona, Modal Sandboxes and 3 more. Price, isolation and pull requests from each vendor, checked 17 October 2026. - [Migrate from Runloop to Rush](https://getrush.ai/compare/migrate-from-runloop.md): How to move work from Runloop to Rush: what to keep, what carries over and the first run. Steps from Runloop AI's own docs, checked 17 October 2026. - [Rush vs Vercel Sandbox (2026)](https://getrush.ai/compare/rush-vs-vercel-sandbox.md): Rush vs Vercel Sandbox: what runs, isolation, pull requests, MCP and price, each cited from Vercel's own pages, checked 17 October 2026. - [Vercel Sandbox alternatives (2026)](https://getrush.ai/compare/vercel-sandbox-alternatives.md): Vercel Sandbox alternatives: Rush, E2B, Daytona, Modal Sandboxes and 3 more. Price, isolation and pull requests from each vendor, checked 17 October 2026. - [Migrate from Vercel Sandbox to Rush](https://getrush.ai/compare/migrate-from-vercel-sandbox.md): How to move work from Vercel Sandbox to Rush: what to keep, what carries over and the first run. Steps from Vercel's own docs, checked 17 October 2026. - [Rush vs Cloudflare Sandbox (2026)](https://getrush.ai/compare/rush-vs-cloudflare-sandbox.md): Rush vs Cloudflare Sandbox: what runs, isolation, pull requests, MCP and price, each cited from Cloudflare's own pages, checked 17 October 2026. - [Cloudflare Sandbox alternatives (2026)](https://getrush.ai/compare/cloudflare-sandbox-alternatives.md): Cloudflare Sandbox alternatives: Rush, E2B, Daytona, Modal Sandboxes and 3 more. Price, isolation and pull requests from each vendor, checked 17 October 2026. - [Migrate from Cloudflare Sandbox to Rush](https://getrush.ai/compare/migrate-from-cloudflare-sandbox.md): How to move work from Cloudflare Sandbox to Rush: what to keep, what carries over and the first run. Steps from Cloudflare's own docs, checked 17 October 2026. - [Rush vs Amazon Bedrock AgentCore (2026)](https://getrush.ai/compare/rush-vs-agentcore.md): Rush vs Amazon Bedrock AgentCore: what runs, isolation, pull requests, MCP and price, each cited from AWS's own pages, checked 17 October 2026. - [Amazon Bedrock AgentCore alternatives (2026)](https://getrush.ai/compare/agentcore-alternatives.md): Amazon Bedrock AgentCore alternatives: Rush, E2B, Daytona and 4 more. Price, isolation and pull requests from each vendor, checked 17 October 2026. - [Rush vs Railway cloud agents (2026)](https://getrush.ai/compare/rush-vs-railway-cloud-agents.md): Rush vs Railway cloud agents: what runs, isolation, pull requests, MCP and price, each cited from Railway's own pages, checked 17 October 2026. - [Railway cloud agents alternatives (2026)](https://getrush.ai/compare/railway-cloud-agents-alternatives.md): Railway cloud agents alternatives: Rush, Claude Code on the web and 5 more. Price, isolation and pull requests from each vendor, checked 15 September 2026. - [Rush vs OpenHands Cloud (2026)](https://getrush.ai/compare/rush-vs-openhands.md): Rush vs OpenHands Cloud: what runs, isolation, pull requests, MCP and price, each cited from All Hands AI's own pages, checked 17 October 2026. - [OpenHands Cloud alternatives (2026)](https://getrush.ai/compare/openhands-alternatives.md): OpenHands Cloud alternatives: Rush, Devin, Tembo, Warp Oz and 3 more. Price, isolation and pull requests from each vendor, checked 15 September 2026. - [Migrate from OpenHands Cloud to Rush](https://getrush.ai/compare/migrate-from-openhands.md): How to move work from OpenHands Cloud to Rush: what to keep, what carries over and the first run. Steps from All Hands AI's own docs, checked 17 October 2026. - [Rush vs Augment Cloud Agents (2026)](https://getrush.ai/compare/rush-vs-augment-cloud-agents.md): Rush vs Augment Cloud Agents: what runs, isolation, pull requests, MCP and price, each cited from Augment Code's own pages, checked 17 October 2026. - [Augment Cloud Agents alternatives (2026)](https://getrush.ai/compare/augment-cloud-agents-alternatives.md): Augment Cloud Agents alternatives: Rush, Cursor Cloud Agents, Devin and 4 more. Price, isolation and pull requests from each vendor, checked 15 September 2026. - [Migrate from Augment Cloud Agents to Rush](https://getrush.ai/compare/migrate-from-augment-cloud-agents.md): How to move work from Augment Cloud Agents to Rush: what to keep, what carries over and the first run. Steps from Augment Code's own docs, checked 17 October 2026. - [Migrate from Ona to Rush](https://getrush.ai/compare/migrate-from-ona.md): How to move work from Ona to Rush: what to keep, what carries over and the first run. Steps from Ona (formerly Gitpod)'s own docs, checked 17 October 2026. - [Rush vs Tembo (2026)](https://getrush.ai/compare/rush-vs-tembo.md): Rush vs Tembo: what runs, isolation, pull requests, MCP and price, each cited from Tembo's own pages, checked 17 October 2026. - [Tembo alternatives (2026)](https://getrush.ai/compare/tembo-alternatives.md): Tembo alternatives: Rush, Devin, OpenHands Cloud, Warp Oz and 3 more. Price, isolation and pull requests from each vendor, checked 15 September 2026. - [Migrate from Tembo to Rush](https://getrush.ai/compare/migrate-from-tembo.md): How to move work from Tembo to Rush: what to keep, what carries over and the first run. Steps from Tembo's own docs, checked 17 October 2026. - [Rush vs Warp Oz (2026)](https://getrush.ai/compare/rush-vs-warp-oz.md): Rush vs Warp Oz: what runs, isolation, pull requests, MCP and price, each cited from Warp's own pages, checked 17 October 2026. - [Warp Oz alternatives (2026)](https://getrush.ai/compare/warp-oz-alternatives.md): Warp Oz alternatives: Rush, Cursor Cloud Agents and 5 more. Price, isolation and pull requests from each vendor, checked 15 September 2026. - [Migrate from Warp Oz to Rush](https://getrush.ai/compare/migrate-from-warp-oz.md): How to move work from Warp Oz to Rush: what to keep, what carries over and the first run. Steps from Warp's own docs, checked 17 October 2026. - [Rush vs CodeRabbit (2026)](https://getrush.ai/compare/rush-vs-coderabbit.md): Rush vs CodeRabbit: what runs, isolation, pull requests, MCP and price, each cited from CodeRabbit's own pages, checked 17 October 2026. - [CodeRabbit alternatives (2026)](https://getrush.ai/compare/coderabbit-alternatives.md): CodeRabbit alternatives: Rush, Greptile, Graphite, Qodo. Price, isolation and pull requests from each vendor, checked 17 October 2026. - [Migrate from CodeRabbit to Rush](https://getrush.ai/compare/migrate-from-coderabbit.md): How to move work from CodeRabbit to Rush: what to keep, what carries over and the first run. Steps from CodeRabbit's own docs, checked 17 October 2026. - [Rush vs Greptile (2026)](https://getrush.ai/compare/rush-vs-greptile.md): Rush vs Greptile: what runs, isolation, pull requests, MCP and price, each cited from Greptile's own pages, checked 17 October 2026. - [Greptile alternatives (2026)](https://getrush.ai/compare/greptile-alternatives.md): Greptile alternatives: Rush, CodeRabbit, Graphite, Qodo. Price, isolation and pull requests from each vendor, checked 17 October 2026. - [Migrate from Greptile to Rush](https://getrush.ai/compare/migrate-from-greptile.md): How to move work from Greptile to Rush: what to keep, what carries over and the first run. Steps from Greptile's own docs, checked 17 October 2026. - [Rush vs Graphite (2026)](https://getrush.ai/compare/rush-vs-graphite.md): Rush vs Graphite: what runs, isolation, pull requests, MCP and price, each cited from Graphite (part of Cursor)'s own pages, checked 17 October 2026. - [Graphite alternatives (2026)](https://getrush.ai/compare/graphite-alternatives.md): Graphite alternatives: Rush, CodeRabbit, Greptile, Qodo. Price, isolation and pull requests from each vendor, checked 17 October 2026. - [Rush vs Replit Agent (2026)](https://getrush.ai/compare/rush-vs-replit.md): Rush vs Replit Agent: what runs, isolation, pull requests, MCP and price, each cited from Replit's own pages, checked 17 October 2026. - [Replit Agent alternatives (2026)](https://getrush.ai/compare/replit-alternatives.md): Replit Agent alternatives: Rush, Lovable, v0, Bolt. Price, isolation and pull requests from each vendor, checked 17 October 2026. - [Migrate from Replit Agent to Rush](https://getrush.ai/compare/migrate-from-replit.md): How to move work from Replit Agent to Rush: what to keep, what carries over and the first run. Steps from Replit's own docs, checked 17 October 2026. - [Lovable alternatives (2026)](https://getrush.ai/compare/lovable-alternatives.md): Lovable alternatives: Rush, Replit Agent, v0, Bolt. Price, isolation and pull requests from each vendor, checked 17 October 2026. - [Migrate from Lovable to Rush](https://getrush.ai/compare/migrate-from-lovable.md): How to move work from Lovable to Rush: what to keep, what carries over and the first run. Steps from Lovable's own docs, checked 17 October 2026. - [Rush vs v0 (2026)](https://getrush.ai/compare/rush-vs-v0.md): Rush vs v0: what runs, isolation, pull requests, MCP and price, each cited from Vercel's own pages, checked 17 October 2026. - [v0 alternatives (2026)](https://getrush.ai/compare/v0-alternatives.md): v0 alternatives: Rush, Replit Agent, Lovable, Bolt. Price, isolation and pull requests from each vendor, checked 17 October 2026. - [E2B vs Daytona (2026)](https://getrush.ai/compare/e2b-vs-daytona.md): E2B vs Daytona: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [E2B vs Modal Sandboxes (2026)](https://getrush.ai/compare/e2b-vs-modal.md): E2B vs Modal Sandboxes: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [Daytona vs Modal Sandboxes (2026)](https://getrush.ai/compare/daytona-vs-modal.md): Daytona vs Modal Sandboxes: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [Vercel Sandbox vs E2B (2026)](https://getrush.ai/compare/vercel-sandbox-vs-e2b.md): Vercel Sandbox vs E2B: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [Vercel Sandbox vs Daytona (2026)](https://getrush.ai/compare/vercel-sandbox-vs-daytona.md): Vercel Sandbox vs Daytona: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [Cloudflare Sandbox vs Vercel Sandbox (2026)](https://getrush.ai/compare/cloudflare-sandbox-vs-vercel-sandbox.md): Cloudflare Sandbox vs Vercel Sandbox: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [E2B vs Cloudflare Sandbox (2026)](https://getrush.ai/compare/e2b-vs-cloudflare-sandbox.md): E2B vs Cloudflare Sandbox: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [Runloop vs E2B (2026)](https://getrush.ai/compare/runloop-vs-e2b.md): Runloop vs E2B: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [Runloop vs Daytona (2026)](https://getrush.ai/compare/runloop-vs-daytona.md): Runloop vs Daytona: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [Amazon Bedrock AgentCore vs E2B (2026)](https://getrush.ai/compare/agentcore-vs-e2b.md): Amazon Bedrock AgentCore vs E2B: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [OpenHands Cloud vs Devin (2026)](https://getrush.ai/compare/openhands-vs-devin.md): OpenHands Cloud vs Devin: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 15 September 2026. - [Tembo vs Devin (2026)](https://getrush.ai/compare/tembo-vs-devin.md): Tembo vs Devin: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 15 September 2026. - [Devin vs GitHub Copilot cloud agent (2026)](https://getrush.ai/compare/devin-vs-github-copilot.md): Devin vs GitHub Copilot cloud agent: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 15 September 2026. - [Devin vs Factory Droids (2026)](https://getrush.ai/compare/devin-vs-factory-droids.md): Devin vs Factory Droids: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 15 September 2026. - [Devin vs Codex cloud (2026)](https://getrush.ai/compare/devin-vs-codex-cloud.md): Devin vs Codex cloud: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 15 September 2026. - [Cursor Cloud Agents vs GitHub Copilot cloud agent (2026)](https://getrush.ai/compare/cursor-cloud-agents-vs-github-copilot.md): Cursor Cloud Agents vs GitHub Copilot cloud agent: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 15 September 2026. - [Claude Code on the web vs Codex cloud (2026)](https://getrush.ai/compare/claude-code-web-vs-codex-cloud.md): Claude Code on the web vs Codex cloud: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 15 September 2026. - [Cursor Cloud Agents vs Claude Code on the web (2026)](https://getrush.ai/compare/cursor-cloud-agents-vs-claude-code-web.md): Cursor Cloud Agents vs Claude Code on the web: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 15 September 2026. - [Codex cloud vs Google Jules (2026)](https://getrush.ai/compare/codex-cloud-vs-jules.md): Codex cloud vs Google Jules: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 15 September 2026. - [GitHub Copilot cloud agent vs Google Jules (2026)](https://getrush.ai/compare/github-copilot-vs-jules.md): GitHub Copilot cloud agent vs Google Jules: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 15 September 2026. - [Warp Oz vs Cursor Cloud Agents (2026)](https://getrush.ai/compare/warp-oz-vs-cursor-cloud-agents.md): Warp Oz vs Cursor Cloud Agents: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 15 September 2026. - [Augment Cloud Agents vs Cursor Cloud Agents (2026)](https://getrush.ai/compare/augment-cloud-agents-vs-cursor-cloud-agents.md): Augment Cloud Agents vs Cursor Cloud Agents: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 15 September 2026. - [OpenHands Cloud vs Tembo (2026)](https://getrush.ai/compare/openhands-vs-tembo.md): OpenHands Cloud vs Tembo: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [CodeRabbit vs Greptile (2026)](https://getrush.ai/compare/coderabbit-vs-greptile.md): CodeRabbit vs Greptile: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [CodeRabbit vs Graphite (2026)](https://getrush.ai/compare/coderabbit-vs-graphite.md): CodeRabbit vs Graphite: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [Greptile vs Graphite (2026)](https://getrush.ai/compare/greptile-vs-graphite.md): Greptile vs Graphite: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [CodeRabbit vs Qodo (2026)](https://getrush.ai/compare/coderabbit-vs-qodo.md): CodeRabbit vs Qodo: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [Replit Agent vs Lovable (2026)](https://getrush.ai/compare/replit-vs-lovable.md): Replit Agent vs Lovable: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [Replit Agent vs Bolt (2026)](https://getrush.ai/compare/replit-vs-bolt.md): Replit Agent vs Bolt: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [Lovable vs Bolt (2026)](https://getrush.ai/compare/lovable-vs-bolt.md): Lovable vs Bolt: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [v0 vs Lovable (2026)](https://getrush.ai/compare/v0-vs-lovable.md): v0 vs Lovable: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. - [v0 vs Replit Agent (2026)](https://getrush.ai/compare/v0-vs-replit.md): v0 vs Replit Agent: what runs, isolation, pull requests, MCP and price from each vendor's pages, checked 17 October 2026. ## Guides - [How to move off Devin](https://getrush.ai/guides/move-from-devin.md): What you keep when you leave Cognition's Devin: git remotes, the session API, the GitHub App, and privacy requests. Then how to run the next Task on Rush. - [How to run Claude Code in the cloud](https://getrush.ai/guides/run-claude-code-in-the-cloud.md): Three runtimes for the same Claude Code harness: Anthropic's web product, Rush, and GitHub's third-party agent preview. Dated 15 September 2026. - [What Is a Multi-Agent System?](https://getrush.ai/guides/multi-agent-systems.md): A multi-agent system is a coordinated group of specialized AI agents that collaborate through delegation, parallel work, and shared context to achieve a goal. ## Learn - [Memory: how Rush remembers](https://getrush.ai/learn/memory-how-rush-remembers.md): Rush separates user memory, agent memory, and session memory so continuity improves without turning every interaction into one giant transcript. - [The artifact loop](https://getrush.ai/learn/the-artifact-loop.md): Why Rush treats reports, drafts, and analyses as first-class objects that can be revised, shared, and handed to the next agent. ## Optional - Full text of the quickstart, every docs page, API operation, error code, use case, guide, lesson and post below, in one file: https://getrush.ai/llms-full.txt - Task pages, one job on one stack each with the exact request a real run was started with, in full: https://getrush.ai/llms-tasks.txt - Blog feed (RSS, full text): https://getrush.ai/blog/rss.xml - [Your Agent Can Use Any App. For a Quarter Fewer Tokens.](https://getrush.ai/blog/computer-cli-osworld.md): computer-cli, the computer-use tool your agent calls, used 26% fewer input tokens than Cua's MCP server on 20 hard OSWorld tasks, at the same pass rate. - [Agents Now Write Their Own Instructions. Who Reviews Them?](https://getrush.ai/blog/agents-that-teach-themselves.md): Agents now keep memory, write their own skills and write the workflows that run other agents. Each of those files changes production, so treat it like code. - [Our Writing Agent Made 858 Commits. Then It Died Quietly.](https://getrush.ai/blog/the-agent-that-died-quietly.md): An unattended agent committed every half hour for 45 days. Then its config vanished and the scheduler ran empty for 64 days, and no alarm went off. - [The Interrupt Is the Interface](https://getrush.ai/blog/intervening-running-coding-agents.md): A running coding agent changes the job. The control that matters is the record that lets you stop, steer, and inspect a run and keep the work. - [Turn Every Repeated Review Comment Into a Failing Test](https://getrush.ai/blog/review-comments-become-tests.md): A style guide is read once; a check fails at the moment of the mistake. A 49-line test took our API from 82 test files mocking our own code to zero in 22 days. - [Meta Gave Every Consumer a Cloud Computer](https://getrush.ai/blog/every-consumer-got-a-cloud-computer.md): Meta Muse runs each user's agent on its own cloud computer, with a second agent approving what leaves and a full record. Consumers got this before most companies. - [The Model Learned CAD. The Spool Still Jammed.](https://getrush.ai/blog/the-model-learned-cad.md): GPT-6 Astra rebuilds parts as CAD code at 95.9% voxel overlap. When our agent printed a part, the CAD was one step of ten, and the checks decided the result. - [We Deleted 862,973 Lines in Nine Days](https://getrush.ai/blog/we-deleted-862973-lines.md): From September 6 to 15 we removed 3.3 lines for every line we added. Agents bias toward adding, so we made deletion the default and made new code argue for itself. - [The Harness Is Free Now](https://getrush.ai/blog/the-harness-is-free-now.md): OpenAI hosts the Codex harness behind an API at no added fee, and AWS runs a managed agent loop. Left to choose: where work runs, what is kept, and the proof. - [The Agent Turned Off Its Own Sandbox](https://getrush.ai/blog/the-agent-turned-off-its-sandbox.md): A confined agent in DeepSeek Harness called its own local control API and switched its session to full access. Isolation the agent can reach, it can undo. - [When You Can't Read the Model's Thoughts, Read Its Actions](https://getrush.ai/blog/read-its-actions.md): GPT-6 Astra writes less of its reasoning down and OpenAI says it is harder to monitor. Oversight moves to what the agent does, and that record is infrastructure. - [Never Let the Author Grade Its Own Work](https://getrush.ai/blog/never-let-the-author-grade-its-work.md): When every agent pushes as one GitHub identity, an agent approving its own pull request looks like a review. A reviewer on another model, and a record, can tell. - [Your Harness Already Has an Eval Set. It's the Transcripts.](https://getrush.ai/blog/measure-the-harness-with-its-transcripts.md): In 30 days our agents started 227 browser tasks, closed 92 and wrote one workaround 1,376 times. Mining our own transcripts found it, with no synthetic benchmark. - [Done Means Installed and Demonstrated](https://getrush.ai/blog/done-means-installed.md): Our terminal crash fix merged and shipped, and sessions still died on the installed build. The finish line is the installed artifact doing the thing, with a capture. - [Instructions Are Read Once. Hooks Fire at the Mistake.](https://getrush.ai/blog/hooks-not-instructions.md): Our rules file told agents in capitals never to build Go binaries by hand, and they kept doing it. A hook that blocks the action and names the next step held. - [One Model, 30.2% and 95.5%. The Harness Was the Difference.](https://getrush.ai/blog/the-harness-gain.md): Prime Intellect moved Claude Opus 5 from 30.2% to 95.5% on ARC-AGI-3 by changing what wraps it. If the gain lives in the harness, the run history is the asset. - [Sixteen Hours at 50%, Three Hours at 80%](https://getrush.ai/blog/sixteen-hours-at-fifty-percent.md): METR measured an agent that finishes sixteen-hour tasks half the time and three-hour tasks four times in five. Checks, retries and proof cover the distance between. - [An Agent You Can't Find Is an Agent You Don't Have](https://getrush.ai/blog/agents-you-cant-find.md): Run more than a handful of agents and the hard engineering stops being intelligence. It becomes addressing: identity, index and resume. The model is rented. - [Your Sandbox Is Only as Isolated as Its Package Proxy](https://getrush.ai/blog/sandbox-only-as-isolated-as-its-proxy.md): OpenAI's evaluation agents reached Hugging Face production through a package registry cache, the one service their sandbox could use. Isolation covers every channel. - [The Fastest Browser Agent Isn't the Cheapest to Run](https://getrush.ai/blog/the-fastest-browser-agent-isnt-the-cheapest-to-run.md): Vercel's agent-browser returns a warm snapshot in 56ms. On a byte-identical page it beat the browser Rush agents drive on latency; ours used 5.8 times fewer tokens. - [The Other Way to Speed Up a Browser Agent](https://getrush.ai/blog/the-other-way-to-speed-up-browser-agents.md): A HAR trick turns a browser task into a typed client, 30 seconds down to 200ms, if the API is stable. The other half: a fast live browser loop, 632ms to 3ms. - [After the Org Chart](https://getrush.ai/blog/after-the-org-chart.md): The model is the steel; your memory is the core. In the agent era everyone rents the same mind, so you cannot out-model your rivals, only out-remember them. - [Every Agent Demo Hides the Same Number](https://getrush.ai/blog/agent-demos-hide-one-number.md): A demo sets the cost of failure to zero and skips independent review. Production restores both. That delta — not the model — is what kills 88% of agent pilots. - [Sovereignty Lives in Your Agents' Trajectories](https://getrush.ai/blog/sovereignty-is-your-trajectories.md): Sovereign AI argues about weights. For a company running agents, the asset is the record of every run, and whoever stores it holds the alpha. - [Loop Engineering Needs Somewhere for the Loops to Run](https://getrush.ai/blog/loops-need-somewhere-to-run.md): When people stop prompting agents and write loops, the loop is infrastructure: a machine off your laptop, isolation per run, a wake-up and a record of each pass. - [The Three-Bet Era](https://getrush.ai/blog/the-three-bet-era.md): Marc Andreessen said software is eating the world, and for fifteen years he was right. The cognition era asks one company for three bets, and wrappers will find out. - [Agents Are Eating the Org Chart](https://getrush.ai/blog/agents-are-eating-the-org-chart.md): AI did not just automate coding. It made organizational judgment loops executable. Software factories are the prototype for a class of programmable institutions.