# Rhino > Rhino is Your agents' cloud for individuals and teams (orgs, teams, > grants): connect the AI agent you already use over MCP and it gains > durable memory, credentials it never reads, saved packages, jobs, > workflows, webhooks, and apps that keep running while you are offline, > and every other agent you connect reuses the same home. Start with [What is Rhino?](https://rhino.nyneelevenlabs.com/docs.md). Each documentation link below is plain markdown; the same content is available to connected agents as `search({ entity: "guide:{id}" })`. ## Kody Open API Call Kody over HTTPS (OpenAPI 3.1). Available to every signed-in account. - [Interactive docs](https://api-docs.kody.codes/) — Scalar UI at api-docs.kody.codes - [OpenAPI description](https://api.kody.codes/openapi.json) — machine-readable service description - [Kody Open API guide](https://rhino.nyneelevenlabs.com/docs/open-api.md) — how to mint tokens and call `/v1` - [API versioning and deprecation](https://rhino.nyneelevenlabs.com/docs/api-versioning.md) — additive `/v1`, phased migrations ## Introduction What Kody is, the two MCP tools, and the loop those tools run. - [What is Kody?](https://rhino.nyneelevenlabs.com/docs/what-is-kody.md) — Start here. Kody is Your agents' cloud for individuals and teams. Connect the AI agent you already use over MCP and it gains durable memory, credentials it never reads, saved packages, and jobs that keep running while you are offline. Covers what Kody is, how it pairs with the agents you already use, what you cannot get elsewhere, the building blocks, real uses, and notes for an agent running a discovery interview. Nothing here requires an account. - [Search and execute](https://rhino.nyneelevenlabs.com/docs/search-and-execute.md) — Search finds capabilities, guides, packages, integrations, and secrets. execute runs an ephemeral module that calls what you found. Covers how they work together and the inputs an agent actually passes. The MCP api tool is documented separately. - [How Kody works](https://rhino.nyneelevenlabs.com/docs/how-kody-works.md) — The factory loop as a conversation: ask what your favorite bot shipped, save the answer as an export you can invoke from any agent, then a daily email that stays quiet until something actually ships. - [The Kody factory map](https://rhino.nyneelevenlabs.com/docs/kody-factory.md) — Map Kody's discovery, credentials, integrations, packages, storage, schedules, hosted surfaces, and memories—and know where the boundary is. ## Get started Connect the agent you already use, make one useful thing, then reuse it from a second agent. - [Connect your agent](https://rhino.nyneelevenlabs.com/docs/connect-your-agent.md) — The three-step Get started flow for people: connect the agent you already use over MCP, paste one prompt so that agent makes something useful in your account, then connect a second agent from another ecosystem and watch it reuse the same memory and package. Includes what approving access means, verified-email requirements, and which agents suit building versus using. - [First run: the onboarding briefing](https://rhino.nyneelevenlabs.com/docs/onboarding.md) — First-run guide for a connected agent. Teach lightly what Kody is and is not, offer six concrete first-win choices, help them take one small win from their pick, and set up integrations with clear difficulty. Then Step 3. - [First build: ad hoc execute, then persist](https://rhino.nyneelevenlabs.com/docs/quick-example.md) — Agent playbook after Make something useful (`/onboarding/step-2`) or Connect your agent (`/onboarding/step-1`): run one useful ad hoc execute, show the result, persist that working code as a package they own, and offer optional triggers without recommending one. - [Second agent: reuse what already exists](https://rhino.nyneelevenlabs.com/docs/portability.md) — Second-agent proof after a first win. Find what the account already has, reuse it once, and stop. Retrieved as guide:portability from Step 3. - [Email and memories: the welcome-email loop](https://rhino.nyneelevenlabs.com/docs/first-win.md) — Optional agent playbook: send a welcome email, point the person at their own inbox to reply, look the reply up when they say it is sent, save what matters as memories, then offer a next step. After Connect your agent (`/onboarding/step-1`), Step 2 is Make something useful (`/onboarding/step-2`) and Step 3 is Connect a second agent (`/onboarding/step-3`); the optional climax guide is quick-example. Load this when someone wants the email loop. Includes the no-polling rule and troubleshooting for missing mail and unbound tools. ## Concepts The primitives every connected agent shares: memory, credentials, packages, triggers, and the runtime. - [Shared memory](https://rhino.nyneelevenlabs.com/docs/memory.md) — What a Kody memory is (personal-only), why it follows the person instead of one host's chat or a team org, how memories surface automatically on search and execute, the verify-first rule for writes, and what does not belong in memory (package state, config, credentials, shared org knowledge). Load this when someone asks what their agents remember about them or how memory travels between agents. - [Where agent guidance lives](https://rhino.nyneelevenlabs.com/docs/agent-guidance.md) — Progressive disclosure for agent guidance across MCP server instructions, package docs (README.md and AGENTS.md), export JSDoc, and memories. Load this when deciding where a fact or instruction should live, or before calling metaMemoryUpsert or metaSetMcpServerInstructions. Search phrases such as "where should this guidance live" should find this page. - [Credentials](https://rhino.nyneelevenlabs.com/docs/credentials.md) — What a Kody credential is and every kind it comes in (Key, OAuth, MCP server, URL, login, signing), how `{{secret:name}}` placeholders and scoped `credential(key)` clients place a value at the network boundary, how packages declare needs in `kody.credentials` (including publisher and executor credential needs that app visitors fill with `credential.visitor(key)`), how callers pick between credentials with names and call maps, the `credential_ambiguous` / `credential_unbound` errors, and webhook verification. Load this when someone asks how credentials work, which kind to use, or why a credential did not resolve. - [Secrets](https://rhino.nyneelevenlabs.com/docs/secrets.md) — How Kody lets your agent use your API keys, tokens, and private keys without ever reading them: the no-secret_get rule, `{{secret:name}}` placeholders resolved at the network boundary, host approval, package approval (website Allow links via `resourceLock`), and expiry. Keys are one credential kind; the Credentials guide covers every kind. Load this when someone asks whether the agent can see a key. - [Custom secret providers](https://rhino.nyneelevenlabs.com/docs/secret-providers.md) — Use credentials from an external vault in secret-aware fetch without the agent ever reading them. - [Packages, integrations, and MCP servers](https://rhino.nyneelevenlabs.com/docs/packages-integrations-mcp.md) — Packages are saved behavior. Integrations are account-owned logins. MCP servers are how agents connect. Use this when those three look interchangeable. - [Organizations](https://rhino.nyneelevenlabs.com/docs/organizations.md) — Personal vs team orgs, invite links, who automation runs as, the org-wide execute pool, connecting one client to several orgs, free-org limits, and paid seats. - [Text your agent](https://rhino.nyneelevenlabs.com/docs/text-your-agent.md) — How to let a spouse or friend talk to *your agent* over a familiar chat surface (iMessage worked example). Surface vs brain vs home: the agent holds the conversation; Kody is the home it reaches into for memory, secrets, packages, and jobs. Covers OpenClaw/Bezalel/Grok Bot patterns and a simpler SMS path with Twilio. - [OpenMuse and Kody](https://rhino.nyneelevenlabs.com/docs/openmuse.md) — How and why to use OpenMuse, CopilotKit's open-source, self-hostable personal agent, alongside Kody. OpenMuse is the personal-agent app (phone and web chat, an agent computer, goals, reviewed actions); Kody is the home that agent shares with your other agents (memory, credentials it never reads, packages, jobs, webhooks, inbox). Covers what each is good at, what to put where, and how to connect OpenMuse's agent harness to Kody over MCP. - [Jobs, workflows, and webhooks](https://rhino.nyneelevenlabs.com/docs/triggers.md) — How work runs when nobody is in the chat. Package-owned jobs for recurring schedules, workflows for deferred or long one-shot runs, inbound webhooks for provider events, subscriptions for events inside Kody, and the inbox — which to pick, how each is declared, and how to test one safely before enabling it. Load this when someone asks how to run something later, on a schedule, or when a provider knocks. - [Runtime and efficiency](https://rhino.nyneelevenlabs.com/docs/platform-efficiency.md) — How unique Dynamic Worker days work across execute, jobs, package exports, and other surfaces, and how the acting user plus a stable module graph reuse one isolate per UTC day. - [Local CLI execute](https://rhino.nyneelevenlabs.com/docs/local-execute.md) — Prefer @kodycodes/cli execute --local when Node ≥22 and the CLI are available. Orthogonal to Open API: use this guide for CLI setup and --local usage; see guide:open_api for HTTPS / MCP api fallback. - [Git credential helper contract (setup-git)](https://rhino.nyneelevenlabs.com/docs/auth-setup-git.md) — Server contract for `kody auth setup-git`: user-owned CLI tokens, `?org=`, and persistent HTTPS git auth with a long-lived `kody_at_…` on `https://kody.codes/@owner/name.git` (clone, fetch, and push). - [Kody Open API](https://rhino.nyneelevenlabs.com/docs/open-api.md) — Call Kody over HTTPS at api.kody.codes. MCP api runs one operation (operationId + params), including tokenCreate / tokenList / tokenRevoke. Available to every signed-in account. - [API versioning and deprecation](https://rhino.nyneelevenlabs.com/docs/api-versioning.md) — How Kody versions the HTTP Open API under /v1, signals deprecations, and migrates without breaking people and their packages. ## Examples Homepage trigger cards as worked examples: cron, webhook, email, and event. - [Flake Hunter](https://rhino.nyneelevenlabs.com/docs/flake-hunter.md) — Homepage cron example: a package-owned daily job that scans CI for flaky end-to-end runs, stays quiet on a clean day, and only acts when a flake signal is present. Load this when someone asks how Flake Hunter works or how to hang a schedule on a package they own. - [Sentry Issues](https://rhino.nyneelevenlabs.com/docs/sentry-issues.md) — Homepage webhook example: a package-owned inbound URL that accepts a Sentry issue POST, verifies HMAC on the platform, and starts triage without a model in the request path. Load this when someone asks how Sentry Issues works or how to hang a vendor webhook on a package they own. - [Agent inbox](https://rhino.nyneelevenlabs.com/docs/agent-inbox.md) — Homepage email example: every account has a platform inbox, inbound mail emits email.message.received, and a package can wake work from a plus-tag without a model reading the mailbox. Load this when someone asks how Agent inbox works or how to treat email as a trigger. - [Purchase thanks](https://rhino.nyneelevenlabs.com/docs/purchase-thanks.md) — Homepage event example: a verified purchase becomes a fact your factory already knows, and a package drafts a thank-you without anyone typing. Load this when someone asks how Purchase thanks works or how to react to an event Kody (or a package you own) already emitted. ## Packages Turn working code into a package you own, then grow it. - [Package lifecycle: reuse, execute, fork, or create](https://rhino.nyneelevenlabs.com/docs/package-lifecycle.md) — Choose between invoking existing behavior, temporary execute exploration, forking a close public package, deferred workflows, and creating a durable repo-backed package; session/tool-only lane validates with `repoRunChecks`, git lane with local checks plus `packagePublishExternalPush`; when the host can run the CLI use `kody dev` / mocks / `--local` before publish. - [Package authoring](https://rhino.nyneelevenlabs.com/docs/package-authoring.md) — START HERE when creating or materially changing a Kody package: package shape, required README.md (human) and AGENTS.md (agent), README Intent section, per-export JSDoc (search Purpose), strict runtime input checking for agent-facing exports, personal-details hygiene before going public, secret-using package approval checklist, and scope-update guidance without adding new primitives. - [Share a package and manage access](https://rhino.nyneelevenlabs.com/docs/package-sharing.md) — Give a person or team Use, Contribute, or Manage on one package with a grant, invite someone who is not in your organization yet, and know what each side can run, read, and pay for. Fork instead when they should own a copy. - [Shared packages for many accounts](https://rhino.nyneelevenlabs.com/docs/shared-packages.md) — Build one package that many callers import, each with its own account of the same provider. Declare the need once, let each caller map it on its own package, pick per run, open several accounts with multiple, or pass a client in. - [Ship Agent Skills in a package](https://rhino.nyneelevenlabs.com/docs/package-skills.md) — Put skills//SKILL.md in a package, publish, and expose them over MCP to hosts that support Skills-over-MCP. - [Package apps](https://rhino.nyneelevenlabs.com/docs/package-apps.md) — Give a package a hosted Worker app on `*.kody.run`. Covers the fetch-handler contract, mount-stripped URLs, `packageContext` / `kody:runtime`, the platform-built browser client (`kody.app.client`) and static assets directory (`kody.app.assets`), session handoff, optional CLI local testing with `kody dev` when available, MCP-only validate via `repoRunChecks`, smoke tests with packageAppFetch, a Remix recipe (Example A), the same-origin proxy, lean forks, compiled clients, and listing verification. - [Custom domains](https://rhino.nyneelevenlabs.com/docs/package-app-custom-domains.md) — Serve an app on a domain you own. Covers which hostnames are allowed, the exact DNS records (DNS only, proxy off), how Kody checks status and emails you, and what pending, active, and needs attention mean. - [Package subscriptions and events](https://rhino.nyneelevenlabs.com/docs/package-subscriptions.md) — Use package.json#kody.subscriptions for package-owned event handlers; discover subscribers with packageSubscriptionsList; smoke-test one handler from interactive MCP with packageSubscriptionDispatch; follow metadata-first email, run.error.recorded activity notifiers, integration.auth.failed / integration.auth.succeeded reconnect notifiers, mcp.server.disconnected / mcp.server.reconnected connection episodes, follow.published / follow.pushed follow notifiers, community.fork.upstream_updated fork notifiers, resource change events (credentials, packages, jobs, triggers, membership, memories), and repo / package lifecycle topics. - [MCP Events from packages](https://rhino.nyneelevenlabs.com/docs/mcp-events.md) — Opt a package topic into MCP Events so ChatGPT (and other webhook-capable MCP clients) can subscribe. Delivery reuses the same package event bus as kody.subscriptions. Behind the mcp-events-extension flag. - [Wake your agent](https://rhino.nyneelevenlabs.com/docs/wake-your-agent.md) — Recipe for waking a connected agent when something happens in Kody (a failed run, a package event, inbound mail). Lead with MCP Events when the MCP client supports the events extension (behind the mcp-events-extension flag; ChatGPT today). Otherwise use a package subscription or job that POSTs to the agent's webhook URL, or the email-wake pattern. Pick by what the client supports. - [Offload work that does not fit a Worker isolate](https://rhino.nyneelevenlabs.com/docs/heavy-work-offload.md) — When package checks or publish rebuilds exceed isolate memory or CPU — usually a large npm graph such as PDF.js — keep the Kody package as a thin orchestrator and run the heavy work in a process the owner operates (Cloudflare Container, Fly machine, or similar). Do not provision paid infra without asking. ## Integrations Bring your own keys, OAuth apps, OpenAPI documents, and MCP servers — and lock them to the code that should use them. - [Integration bootstrap](https://rhino.nyneelevenlabs.com/docs/integration-bootstrap.md) — START HERE when a third-party integration must work before saving a dependent package or package app: inspect integration/secret state, stop for setup, run an authenticated smoke test, then communitySearch before building from scratch. - [OAuth: bring your own app](https://rhino.nyneelevenlabs.com/docs/oauth.md) — START HERE for third-party OAuth: hosted /connect/oauth, the exact redirect URI (https://kody.codes/connect/oauth), required query params, PKCE vs confidential, the post-connect copyable agent prompt, and how it differs from MCP OAuth. - [Google OAuth walkthrough](https://rhino.nyneelevenlabs.com/docs/google-oauth.md) — Interactive transcript of a coding agent walking a naive user through Google OAuth for Gmail inbox reading: official guides first, one console step at a time, prefilled /connect/oauth, refuse secrets in chat, then a Gmail profile smoke test. - [Secret-backed integrations](https://rhino.nyneelevenlabs.com/docs/secret-backed-integration.md) — Default non-OAuth recipe for secret-backed integrations: research auth, collect secrets through /connect/secret-set, run a smoke test, build the downstream package, then surface package secret approval links (prefer bulk) before calling the work complete. - [Secret setup URL reference](https://rhino.nyneelevenlabs.com/docs/account-secret-setup.md) — Hosted /connect/secret-set URL shape, query params, approval policy for API keys and PATs, and post-hoc package approval URLs including bulk approve. - [OpenAPI integrations](https://rhino.nyneelevenlabs.com/docs/openapi-integrations.md) — Prefer a close community helpers package, then fork `@kody/openapi` to bind and call selected operations. - [Connect a home MCP server](https://rhino.nyneelevenlabs.com/docs/local-mcp-tunnels.md) — Run an MCP server beside a vault, CLI, or home device, publish it with Cloudflare Tunnel and Access, and connect its HTTPS URL to Kody. - [Lock an MCP server to a package](https://rhino.nyneelevenlabs.com/docs/locked-mcp-server.md) — Connected MCP servers expose every discovered tool to execute and every package. Lock the server to a thin package so only that published surface can call it. - [Gmail drafts without send](https://rhino.nyneelevenlabs.com/docs/locked-gmail-drafts.md) — Google has gmail.send but no drafts-only scope. gmail.compose can manage drafts and send. Publish a drafts-only package, lock it, then resourceLock the connection so execute cannot use the token. ## Connect a provider Verified, console-by-console walkthroughs for connecting a specific service. - [Connect Discord](https://rhino.nyneelevenlabs.com/docs/discord.md) — Verified walkthrough for connecting Discord to Kody: the OAuth user-token lane for identity and server membership, the bot-token lane required for reading or posting in servers, 7-day rotating tokens, prefilled connect links, and users/@me smoke tests for both lanes. - [Connect Dropbox](https://rhino.nyneelevenlabs.com/docs/dropbox.md) — Walkthrough for connecting Dropbox to Kody with your own Dropbox app: app console setup, the PKCE connect link with offline refresh tokens, the three Dropbox API hosts (api, content, notify), and a smoke test through @kentcdodds/dropbox. - [Connect Figma](https://rhino.nyneelevenlabs.com/docs/figma.md) — Verified walkthrough for connecting Figma to Kody (BYO OAuth or personal/plan access token), then using @kody/figma over Figma REST, not Figma's remote MCP. - [Connect GitHub](https://rhino.nyneelevenlabs.com/docs/github.md) — Verified walkthrough for connecting GitHub to Kody: fine-grained personal access tokens, bring-your-own OAuth Apps, scope choices, prefilled connect links, and copy-paste smoke tests. - [Connect Google (Gmail, Calendar, Drive)](https://rhino.nyneelevenlabs.com/docs/google.md) — Verified walkthrough for connecting Google to Kody: bring-your-own OAuth client, Google Auth Platform console steps, the Testing-status 7-day refresh-token trap, and a Calendar or Gmail smoke test. - [Connect Notion](https://rhino.nyneelevenlabs.com/docs/notion.md) — Verified walkthrough for connecting Notion to Kody: the personal internal connection token lane (fastest), the public connection OAuth lane, the page-picker access model (Notion has no OAuth scopes), prefilled connect links, and a users/me smoke test for both lanes. - [Connect Origin](https://rhino.nyneelevenlabs.com/docs/origin.md) — Verified walkthrough for connecting Cursor Origin to Kody: Origin Apps with Ed25519 signing keys, installation access tokens, the community helpers package, and a rate-limit smoke test. Cloud Agents API keys are not Origin credentials. - [Connect Salesforce](https://rhino.nyneelevenlabs.com/docs/salesforce.md) — Verified connect walkthrough for Salesforce: bring-your-own Connected App or External Client App, the exact /connect/oauth callback, production vs sandbox login hosts, scopes, prefilled connect links, and a smoke test. - [Connect Slack](https://rhino.nyneelevenlabs.com/docs/slack.md) — Verified walkthrough for connecting Slack to Kody with a user-token Slack app, then calling it from `@kody/slack`. - [Connect Spotify](https://rhino.nyneelevenlabs.com/docs/spotify.md) — Verified walkthrough for connecting Spotify to Kody: dashboard app creation, the Development Mode Premium requirement and endpoint limits in force since February 2026, the PKCE public-client connect link, listening and playback scopes, and a profile smoke test. ## Help When Kody gets in the way. - [If you can't open kody.codes or get mail from us](https://rhino.nyneelevenlabs.com/docs/cant-open-or-get-mail.md) — Unblock NextDNS and similar DNS filters for kody.codes, and recover missing signup or verification mail when a provider is blocking our domain. - [Report friction](https://rhino.nyneelevenlabs.com/docs/platform-friction.md) — Use for meaningful Kody friction, bugs, poor experiences, or suggestions: show the exact proposal and account-identity notification disclosure before asking for explicit approval, and write one actionable issue per submission (repro steps, expected vs actual, impact). ## Website - [What will you build with Rhino?](https://rhino.nyneelevenlabs.com/use-cases): Build reusable tools, automate inbox work, share context across your agents, and run workflows on a schedule with Kody. - [n8n alternatives for developers](https://rhino.nyneelevenlabs.com/compare/n8n-alternatives): Build reusable workflows with your agent in Kody. Compare its packages, connected accounts, and schedules with n8n and other automation tools. - [An MCP gateway for your personal agents](https://rhino.nyneelevenlabs.com/use-cases/mcp-gateway): Understand direct MCP connections, gateways, and Kody. Connect agents to shared tools, accounts, memory, and saved workflows. - [Shared memory for Claude Code, Cursor, and Codex](https://rhino.nyneelevenlabs.com/use-cases/shared-agent-memory): Share useful facts between Claude Code, Cursor, and Codex. Keep repeatable behavior in packages, and use Kody memory for facts without a better home. - [Gmail automation with Claude and MCP](https://rhino.nyneelevenlabs.com/integrations/gmail): Connect Gmail to your agent and build an inbox workflow you can use again, with drafts to review and control over what gets sent. - [Build reusable custom tools for Claude Code](https://rhino.nyneelevenlabs.com/use-cases/claude-code-custom-tools): Turn a repeated Claude Code task into a Kody package. Keep the code and connected accounts ready for your other agents to use. - [Composio alternatives for personal agent workflows](https://rhino.nyneelevenlabs.com/compare/composio-alternatives): Explore Kody as a Composio alternative for personal agent workflows, with connected accounts, reusable packages, memory, and schedules in one place. - [Slack automation that keeps track of the work](https://rhino.nyneelevenlabs.com/integrations/slack): Build reusable Slack workflows with Kody. Connect your services, keep track of what’s been posted, and review messages before putting them on a schedule. - [Schedule reusable workflows built with Claude](https://rhino.nyneelevenlabs.com/use-cases/scheduled-workflows): Run saved code on a schedule with Kody. Automate reports, keep track of progress, and use AI where your workflow needs it. - [Claude integrations you can keep using across agents](https://rhino.nyneelevenlabs.com/integrations/claude): Connect Claude to Kody and reuse your accounts, memory, and tools beyond one conversation. Build workflows you can run from any connected agent. - [AI workflow automation for developers](https://rhino.nyneelevenlabs.com/use-cases/ai-workflow-automation): Choose visual workflows, agent routines, or reusable code. Build automation in Kody with connected accounts, saved packages, memory, and schedules. - [The cloud for your team’s agents](https://rhino.nyneelevenlabs.com/): One place for the credentials, tools, and workflows for every agent your team members use. - [Public packages — Rhino](https://rhino.nyneelevenlabs.com/community): Browse public packages shared by Kody users. - [Blog — Rhino](https://rhino.nyneelevenlabs.com/blog): Updates and positioning posts from Kent C. Dodds about Kody. - [Rhino features for your AI agents | Rhino](https://rhino.nyneelevenlabs.com/features): Memory, packages, triggers, credentials, and apps for your connected AI agents. - [Shared memory for your AI agents | Rhino](https://rhino.nyneelevenlabs.com/features/memory): Save facts and preferences in Kody and use them with agents connected to your account. - [Reusable packages for your AI agents | Rhino](https://rhino.nyneelevenlabs.com/features/packages): Save reusable code, review its source, publish changes, and share access to your packages. - [Triggers and automation for your AI agents | Rhino](https://rhino.nyneelevenlabs.com/features/triggers): Start saved packages from schedules, webhooks, incoming email, and package events. - [Credentials for your AI agents | Rhino](https://rhino.nyneelevenlabs.com/features/credentials): Keep API keys and OAuth connections out of the conversation. Anything that speaks OAuth can connect, and Kody sends credentials only to hosts you approve. - [Package apps for your AI agents | Rhino](https://rhino.nyneelevenlabs.com/features/apps): Build browser interfaces backed by saved packages and durable package storage. - [AI Agent Cloud for Business and Agencies | Rhino](https://rhino.nyneelevenlabs.com/for/business): Give your teams and AI agents a shared home for tools, data, and automations. Run work across departments and client organizations with clear ownership. - [Pricing — Rhino](https://rhino.nyneelevenlabs.com/pricing): Every plan is the whole factory. You pay for volume. Paid plans raise the caps. - [FAQ — Rhino](https://rhino.nyneelevenlabs.com/faq): Common questions about what Kody is, what it is not, and how your assistant stays yours. - [Case studies — Rhino](https://rhino.nyneelevenlabs.com/case-studies): Longer notes from people using Kody — how it shows up in their work. - [What we believe: Rhino](https://rhino.nyneelevenlabs.com/believe): The beliefs behind Kody, each tied to how the product behaves today. - [Support — Rhino](https://rhino.nyneelevenlabs.com/support): Contact support for the hosted Kody service at kody.codes. - [About Rhino](https://rhino.nyneelevenlabs.com/about): Kody is Your agents' cloud for individuals and teams. Built by Kent C. Dodds at kody.codes. - [Brand guide — Rhino](https://rhino.nyneelevenlabs.com/brand-guide): Download the Kody koala mark as SVG or PNG, and see light and dark treatments. - [Contact Rhino](https://rhino.nyneelevenlabs.com/contact): Email support@kody.codes, join Discord, or open a GitHub issue for the hosted Kody service. - [Privacy — Rhino](https://rhino.nyneelevenlabs.com/privacy): How Kody stores your data and what a deployment admin can see. - [Terms — Rhino](https://rhino.nyneelevenlabs.com/terms): Terms and acceptable use for this Kody deployment. - [Discord — Rhino](https://rhino.nyneelevenlabs.com/discord): Join the official Kody Discord and connect your account so we can assign member and plan roles. - [Get started — Rhino](https://rhino.nyneelevenlabs.com/onboarding): Connect your AI agent host and set up Kody for you or your team. - [Rhino for Cursor](https://rhino.nyneelevenlabs.com/with/cursor): Keep building in Cursor. Kody gives it your team’s credentials, packages, and workflows, and anything it saves runs from every other agent your team uses. - [Rhino for Claude Code](https://rhino.nyneelevenlabs.com/with/claude-code): Claude Code writes the package. Kody keeps it, runs it on a trigger, and shares it with the rest of your team. - [Rhino for Codex](https://rhino.nyneelevenlabs.com/with/codex): Let Codex build tools your whole team can run, with credentials it uses but never sees. - [Rhino for ChatGPT](https://rhino.nyneelevenlabs.com/with/chatgpt): Use your team’s packages and connected accounts from a ChatGPT conversation, on the web or on your phone. - [Rhino for Claude](https://rhino.nyneelevenlabs.com/with/claude): Bring your team’s tools and connected accounts into Claude conversations on desktop, web, and mobile. - [Rhino for GitHub Copilot](https://rhino.nyneelevenlabs.com/with/copilot): Give Copilot in VS Code or the terminal the same credentials and packages the rest of your team uses. - [Rhino for Devin](https://rhino.nyneelevenlabs.com/with/devin): Let Devin use your team’s credentials and packages without pasting a single key into a session. - [Rhino for Gemini](https://rhino.nyneelevenlabs.com/with/gemini): Run your team’s packages from the Gemini app, with credentials Gemini never sees. - [Rhino for Grok](https://rhino.nyneelevenlabs.com/with/grok): Bring your team’s tools into Grok conversations through one custom connector. - [Rhino for Grok Bot](https://rhino.nyneelevenlabs.com/with/grok-bot): Start from the Kody Bot template and Grok Bot works on your Kody repos and packages through the Kody CLI and git. - [Rhino for OpenCode](https://rhino.nyneelevenlabs.com/with/opencode): Point OpenCode at Kody, and every package it writes becomes something your whole team can run. - [Rhino for OpenClaw](https://rhino.nyneelevenlabs.com/with/openclaw): Give your local-first OpenClaw agent the shared credentials and workflows your team keeps in Kody. - [Rhino for Hermes Agent](https://rhino.nyneelevenlabs.com/with/hermes): Give Hermes the shared credentials and packages your team keeps in Kody, from the terminal or any chat app it runs in. - [Rhino for Muse](https://rhino.nyneelevenlabs.com/with/muse): Connect Muse Code to Kody so the tools it builds outlive the session and reach the rest of your team. - [Rhino for Amp](https://rhino.nyneelevenlabs.com/with/amp): Let Amp build packages your whole team can run, with credentials it uses but never sees. - [Rhino for Warp](https://rhino.nyneelevenlabs.com/with/warp): Give Warp’s agents the same credentials, packages, and workflows the rest of your team uses. - [Rhino for Goose](https://rhino.nyneelevenlabs.com/with/goose): Point Goose at Kody, and the tools it builds stay around for the rest of your team. - [Rhino for Zed](https://rhino.nyneelevenlabs.com/with/zed): Bring your team’s credentials, packages, and workflows into Zed’s agent panel. - [Rhino for Pi](https://rhino.nyneelevenlabs.com/with/pi): Give Pi the shared credentials and workflows your team keeps in Kody. - [Rhino for Wajo](https://rhino.nyneelevenlabs.com/with/wajo): Reach your team’s packages and connected accounts from Wajo, wherever you chat. - [Rhino for Cue](https://rhino.nyneelevenlabs.com/with/cue): Let your Cue agents use your team’s tools and connected accounts without ever seeing a key. - [Rhino for OpenMuse](https://rhino.nyneelevenlabs.com/with/openmuse): Wire Kody into your OpenMuse harness so the work it does reaches the rest of your team. - [Rhino for Dots](https://rhino.nyneelevenlabs.com/with/dots): Bring your team’s tools and connected accounts into Dots through one MCP connection. - [Rhino Docs](https://rhino.nyneelevenlabs.com/docs): Learn how to use Rhino. - [Connect a provider](https://rhino.nyneelevenlabs.com/docs/connect): Set up connections to your accounts. - [Connect your agent](https://rhino.nyneelevenlabs.com/auth.md): OAuth and MCP endpoint details. ## Blog posts - [Case studies](https://rhino.nyneelevenlabs.com/blog/early-kody-users): Longer notes from people using Kody — how it shows up in their work, in their words. - [OpenClaw 2 needs a home](https://rhino.nyneelevenlabs.com/blog/openclaw-2-needs-a-home): OpenClaw 2.0 is a personal AI that lives on your machine. Kody is the runtime that keeps its memory, keys, packages, and jobs running when that machine sleeps. Here's how they stack. - [How to turn agent work into software you own](https://rhino.nyneelevenlabs.com/blog/how-to-turn-agent-work-into-software-you-own): Ask once in the agent you already use, persist the working code as a package you own, then trigger it — invoke, job, webhook, or app — without paying the model again. - [Your agents' cloud](https://rhino.nyneelevenlabs.com/blog/your-assistants-home): Your AI assistant accumulates things worth keeping. Memory, keys, code, scheduled jobs, an inbox. Those things should live somewhere durable that you own, no matter which agent you talk to this year. - [Gateways connect. Homes accumulate.](https://rhino.nyneelevenlabs.com/blog/gateways-connect-homes-accumulate): MCP gateways and Kody get compared a lot, but they solve different problems. A gateway is a catalog of tools with pluggable ways to call them. Kody is a runtime where your assistant's stuff lives and keeps running. The difference shows up the moment you try to toggle code mode off. - [The personal assistant that can't leak your keychain](https://rhino.nyneelevenlabs.com/blog/the-assistant-that-cant-leak-your-keychain): My assistant's account holds 51 secrets. Here's the security architecture that makes the scary failure modes structurally impossible instead of policy-discouraged. - [Why Your agents' cloud should make zero inference calls](https://rhino.nyneelevenlabs.com/blog/zero-inference-calls): Kody makes zero inference calls. That's not a limitation, it's the business model, and it changes who your assistant platform is actually working for. - [Every install is a fork you own](https://rhino.nyneelevenlabs.com/blog/every-install-is-a-fork-you-own): In most ecosystems, installing someone's automation means renting a black box. In Kody's community, one-click install creates a fork you own. Here's why that changes how you treat other people's code. - [The automations you never got around to building](https://rhino.nyneelevenlabs.com/blog/the-automations-you-never-built): Your real competitor for automation isn't a better tool, it's not automating at all. Why the activation energy of hosting and credentials killed your automations, and what changes when your agent has a durable home. - [Self-service means ownership](https://rhino.nyneelevenlabs.com/blog/self-service-means-ownership): Why Kody makes you create your own OAuth apps and bring your own keys, framed honestly. A few minutes of setup buy you your own app, your scopes, your rate limits, and no middleman between you and your data.