# ViteHub

> Portable Agents and Server Primitives for any Vite host.

## Documentation Sets

- [ViteHub](https://vitehub.dev/llms-full.txt): Server primitives for Vite.

## When to use ViteHub

Use ViteHub when a Vite application needs portable server storage, queues, workflows, schedules, sandboxes, or other Server Primitives; when it needs inspectable Agent Definitions with explicit Capabilities and Workspaces; or when the same runtime behavior must deploy across supported hosts without leaking provider APIs into application code.

- [Choose a ViteHub layer](https://vitehub.dev/raw/docs/getting-started.md): Decide whether the job needs a Server Primitive, an Agent, or both.
- [Install ViteHub](https://vitehub.dev/raw/docs/getting-started/installation.md): Install the framework distribution or a focused owner package.

## ViteHub developer resources

Use these named machine interfaces to inspect ViteHub before reading the full documentation set. ViteHub is installed into your application and does not expose a shared hosted runtime API.

- [ViteHub OpenAPI document](https://vitehub.dev/openapi.json): Discover the public machine-readable resources served by vitehub.dev.
- [ViteHub Agent Skill](https://vitehub.dev/.well-known/skills/vitehub/SKILL.md): Install or inspect the coding-agent instructions and routed references.
- [ViteHub MCP server](https://vitehub.dev/mcp): Connect an MCP client to the Streamable HTTP documentation endpoint.
- [ViteHub CLI on npm](https://www.npmjs.com/package/vite-hub): Install the official CLI included with the vite-hub package.

## Project and trust information

- [About ViteHub](https://vitehub.dev/raw/about.md)
- [Contact ViteHub](https://vitehub.dev/raw/contact.md)
- [ViteHub privacy](https://vitehub.dev/raw/privacy.md)

## Docs

- [Agent Actors](https://vitehub.dev/raw/docs/agents/actors.md): Carry trusted caller identity into one Agent Invocation.
- [Agent Definitions](https://vitehub.dev/raw/docs/agents/agent-definitions.md): Declare how one Agent runs, what it can use, and how callers reach it.
- [Agent Drivers](https://vitehub.dev/raw/docs/agents/agent-drivers.md): Choose model-backed, provider-backed, or application-owned Agent execution.
- [Boxes](https://vitehub.dev/raw/docs/agents/boxes.md): Understand when application code uses a Box and why it does not attach to an Agent Definition.
- [Channels](https://vitehub.dev/raw/docs/agents/channels.md): Connect an Agent to web chat and messaging platforms without mixing transport with identity.
- [Chat History and sessions](https://vitehub.dev/raw/docs/agents/chat-history-sessions.md): Select prior conversation messages without confusing them with durable Agent Memory.
- [Child invocations](https://vitehub.dev/raw/docs/agents/controlled-child-invocations.md): Start, inspect, respond to, and cancel child Agent work from trusted code.
- [Evals](https://vitehub.dev/raw/docs/agents/evals.md): Run repeatable scenarios against an Agent Definition and score its behavior.
- [Agents](https://vitehub.dev/raw/docs/agents.md): Define a server-side Agent, choose how it runs, and connect it to your application.
- [Instructions](https://vitehub.dev/raw/docs/agents/instructions.md): Write durable model-facing behavior and policy for an Agent.
- [Invocations](https://vitehub.dev/raw/docs/agents/invocations.md): Run, stream, and observe one request to an Agent.
- [Triggers](https://vitehub.dev/raw/docs/agents/triggers.md): Translate product events into Agent Invocations while the Driver continues to own execution.
- [Workspace context](https://vitehub.dev/raw/docs/agents/workspace-context.md): Give an Agent scoped files and Sources with explicit read and write authority.
- [ViteHub coding-agent skill](https://vitehub.dev/raw/docs/ai-resources/agent-instructions-skills.md): Install the public ViteHub skill for Cursor, Claude Code, Codex, and other coding agents.
- [AI-readable documentation](https://vitehub.dev/raw/docs/ai-resources.md): Give coding agents the smallest reliable ViteHub context for the task.
- [Raw Markdown pages](https://vitehub.dev/raw/docs/ai-resources/markdown-pages.md): Load one canonical ViteHub documentation page as clean Markdown.
- [Access](https://vitehub.dev/raw/docs/capabilities/access.md): Resolve trusted invocation access before other Capabilities add tools or Workspace access.
- [Blob](https://vitehub.dev/raw/docs/capabilities/blob.md): Expose scoped Blob read and optional edit tools to an Agent.
- [Browser](https://vitehub.dev/raw/docs/capabilities/browser.md): Give a Provider Agent headless browser guidance.
- [Chat](https://vitehub.dev/raw/docs/capabilities/chat.md): Add chat-oriented Agent Trigger behavior and Chat History state requirements.
- [Chat summary](https://vitehub.dev/raw/docs/capabilities/chat-summary.md): Add a summary command that replaces explicit input with a conversation summary.
- [Cost](https://vitehub.dev/raw/docs/capabilities/cost.md): Enrich Agent Usage Records with exact, display-ready USD cost.
- [Custom capabilities](https://vitehub.dev/raw/docs/capabilities/custom-capabilities.md): Build one product-specific Agent ability with requirements, tools, policy, and metadata.
- [Database](https://vitehub.dev/raw/docs/capabilities/db.md): Expose guarded database schema, query, and mutation tools to an Agent.
- [Diagnostics](https://vitehub.dev/raw/docs/capabilities/diagnostics.md): Report Agent Invocation outcomes and scoped runtime resource observations.
- [Email](https://vitehub.dev/raw/docs/capabilities/email.md): Let an Agent send authorized plain-text email through the configured Email primitive.
- [Fetch](https://vitehub.dev/raw/docs/capabilities/fetch.md): Expose named HTTP request tools for developer-approved endpoints.
- [Git](https://vitehub.dev/raw/docs/capabilities/git.md): Give an Agent bounded Git source-history access inside a Workspace Session.
- [Gmail](https://vitehub.dev/raw/docs/capabilities/gmail.md): Let an Agent search Gmail and create unsent drafts through structured tools.
- [Capabilities](https://vitehub.dev/raw/docs/capabilities.md): Give an Agent tools and behavior without handing it unrestricted server access.
- [Input commands](https://vitehub.dev/raw/docs/capabilities/input-commands.md): Transform explicit user commands before the main Agent Invocation runs.
- [KV](https://vitehub.dev/raw/docs/capabilities/kv.md): Expose scoped KV read and optional edit tools to an Agent.
- [LLM gate](https://vitehub.dev/raw/docs/capabilities/llm-gate.md): Allow or reject an Agent Invocation with a pre-invocation model decision.
- [LLM route](https://vitehub.dev/raw/docs/capabilities/llm-route.md): Choose one developer-defined route before the main Agent Invocation runs.
- [MCP](https://vitehub.dev/raw/docs/capabilities/mcp.md): Connect an Agent to external MCP server tools.
- [Memory](https://vitehub.dev/raw/docs/capabilities/memory.md): Add scoped durable memory records to an Agent.
- [Official capabilities](https://vitehub.dev/raw/docs/capabilities/official-capabilities.md): Choose an official Capability by what the Agent needs to do.
- [OpenAPI](https://vitehub.dev/raw/docs/capabilities/openapi.md): Expose selected OpenAPI operations as bounded Agent tools or a generated Capability CLI.
- [OTLP](https://vitehub.dev/raw/docs/capabilities/otlp.md): Export live Agent Invocation events and completed traces to an OpenTelemetry receiver.
- [Papercuts](https://vitehub.dev/raw/docs/capabilities/papercuts.md): Let an Agent report small runtime and developer-experience friction to an application-owned sink.
- [Progress summary](https://vitehub.dev/raw/docs/capabilities/progress-summary.md): Stream a short user-facing summary of current Agent activity.
- [Rate limit](https://vitehub.dev/raw/docs/capabilities/rate-limit.md): Consume a Rate Limiter before an Agent Invocation starts.
- [Repository host](https://vitehub.dev/raw/docs/capabilities/repository-host.md): Expose provider-hosted repository, Change Request, issue, comment, check, and status data.
- [Repository host context](https://vitehub.dev/raw/docs/capabilities/repository-host-context.md): Read issue and Change Request context through a lazy async record.
- [Sandbox](https://vitehub.dev/raw/docs/capabilities/sandbox.md): Let an Agent run allowlisted executables in an isolated sandbox.
- [Schedule](https://vitehub.dev/raw/docs/capabilities/schedule.md): Declare Agent Schedules or let an Agent manage Runtime Schedules through one cronjob tool.
- [Skills](https://vitehub.dev/raw/docs/capabilities/skills.md): Mount skill files into an Agent Workspace.
- [Subagents](https://vitehub.dev/raw/docs/capabilities/subagents.md): Delegate bounded work to named Agent Definitions through model-facing tools.
- [Title](https://vitehub.dev/raw/docs/capabilities/title.md): Generate a short title and attach it to Agent output.
- [Transcribe](https://vitehub.dev/raw/docs/capabilities/transcribe.md): Convert audio input parts into transcript text before an Agent runs.
- [Usage](https://vitehub.dev/raw/docs/capabilities/usage.md): Request provider usage metadata and expose the normalized Agent Usage Record at finish.
- [Web search](https://vitehub.dev/raw/docs/capabilities/web-search.md): Add model web search or normalized web search and read tools.
- [Workspace shell](https://vitehub.dev/raw/docs/capabilities/workspace-shell.md): Expose scoped Workspace inspection, mutation, and explicit command tools.
- [Agent Invocations](https://vitehub.dev/raw/docs/concepts/agent-invocations.md): Understand what ViteHub resolves and records for one Agent request.
- [Auth Users and Agent Invokers](https://vitehub.dev/raw/docs/concepts/auth-users-and-agent-invokers.md): Understand how application identity becomes trusted invocation identity.
- [Bash](https://vitehub.dev/raw/docs/concepts/bash.md): Understand how Capability commands become one constrained Agent tool.
- [Capabilities](https://vitehub.dev/raw/docs/concepts/capabilities-api.md): Understand how an Agent receives a selected ability.
- [Channels](https://vitehub.dev/raw/docs/concepts/channels-api.md): Understand how messages reach an Agent Invocation.
- [Definition discovery](https://vitehub.dev/raw/docs/concepts/definitions-and-discovery.md): Understand how a ViteHub file becomes a named runtime entry.
- [Concepts](https://vitehub.dev/raw/docs/concepts.md): Learn the ViteHub terms that help you choose an API and understand what runs where.
- [Runtime Context](https://vitehub.dev/raw/docs/concepts/runtime-context.md): Understand the host resources passed into a server operation or Agent request.
- [Runtime Helpers and stable imports](https://vitehub.dev/raw/docs/concepts/runtime-helpers-and-stable-imports.md): Understand which imports application code uses to call ViteHub.
- [Runtime policy, approvals, and traces](https://vitehub.dev/raw/docs/concepts/runtime-policy-approvals-and-traces.md): Understand the records that explain whether runtime work ran, waited, or failed.
- [Server primitives](https://vitehub.dev/raw/docs/concepts/server-primitives-for-any-host.md): Build server-backed features with databases, queues, storage, and more while keeping your application portable across hosts.
- [Vite integrations and Provider Output](https://vitehub.dev/raw/docs/concepts/vite-integrations-and-provider-output.md): Understand how ViteHub prepares a package for development and deployment.
- [Workspace and Sources](https://vitehub.dev/raw/docs/concepts/workspace-and-sources.md): Understand how persistent files differ from read-only origins.
- [CLI](https://vitehub.dev/raw/docs/development/cli.md): Run package-owned development workflows through the local Vite config.
- [Console](https://vitehub.dev/raw/docs/development/console.md): Enable the Agent invocation console, protect its routes, and understand its storage limits.
- [Generated files](https://vitehub.dev/raw/docs/development/generated-files.md): Inspect ViteHub generated files without making them public authoring surfaces.
- [Local development](https://vitehub.dev/raw/docs/development.md): Run ViteHub locally, inspect generated files, and verify the behavior you changed.
- [Provisioning](https://vitehub.dev/raw/docs/development/provisioning.md): Create missing provider resources and write non-secret provider ids into local provision state.
- [Troubleshooting](https://vitehub.dev/raw/docs/development/troubleshooting.md): Diagnose common ViteHub development failures from symptoms to proof paths.
- [Verification](https://vitehub.dev/raw/docs/development/verification.md): Choose the right proof tier for ViteHub primitives and provider behavior.
- [Cloudflare](https://vitehub.dev/raw/docs/frameworks-hosts/cloudflare.md): Configure Cloudflare Provider Output while keeping ViteHub Definitions and Runtime Helpers host-neutral.
- [Deno](https://vitehub.dev/raw/docs/frameworks-hosts/deno.md): Generate Deno server output for Agent routes and Schedule wake output without making app code Deno-specific.
- [Frameworks and hosts](https://vitehub.dev/raw/docs/frameworks-hosts.md): See what ViteHub configures for each framework and deployment host.
- [Netlify](https://vitehub.dev/raw/docs/frameworks-hosts/netlify.md): Use the ViteHub package contracts that have explicit Netlify runtime or function output.
- [Nitro and UnJS](https://vitehub.dev/raw/docs/frameworks-hosts/nitro-unjs.md): Understand the narrow boundary between ViteHub, Nitro, and UnJS server runtimes.
- [Node and self-hosted](https://vitehub.dev/raw/docs/frameworks-hosts/node-self-hosted.md): Use ViteHub Runtime Helpers in Node-shaped hosts without pretending every primitive has unified self-hosted output.
- [Runtime and host support](https://vitehub.dev/raw/docs/frameworks-hosts/support-matrix.md): Check which ViteHub contracts exist for each host and how the repository proves them.
- [Vercel](https://vitehub.dev/raw/docs/frameworks-hosts/vercel.md): Generate Vercel Provider Output while preserving ViteHub package boundaries.
- [First Agent](https://vitehub.dev/raw/docs/getting-started/first-agent.md): Define a server-side Agent, call it from H3, and see the response.
- [First Server Primitive](https://vitehub.dev/raw/docs/getting-started/first-server-primitive.md): Add local KV to a small Vite server and return one stored value.
- [Introduction](https://vitehub.dev/raw/docs/getting-started.md): Choose the ViteHub layer that matches the first thing you need to build.
- [Installation](https://vitehub.dev/raw/docs/getting-started/installation.md): Install the ViteHub framework distribution or choose a direct owner package for advanced composition.
- [ViteHub docs](https://vitehub.dev/raw/docs.md): Add server APIs or portable Agents to a Vite application.
- [Channels](https://vitehub.dev/raw/docs/reference/channels.md): Define named server delivery channels and send through a selected connector.
- [Config options](https://vitehub.dev/raw/docs/reference/config-options.md): Reference the main ViteHub Integration Options and where provider choices belong.
- [Errors and diagnostics](https://vitehub.dev/raw/docs/reference/errors-diagnostics.md): Reference ViteHub error codes and the local proof path for each primitive.
- [File conventions](https://vitehub.dev/raw/docs/reference/file-conventions.md): Reference the discovery paths that produce ViteHub Discovered Definitions and Discovery Identity.
- [Import paths](https://vitehub.dev/raw/docs/reference/import-paths.md): Distinguish stable ViteHub imports from generated files, provider modules, and internal package paths.
- [Package reference](https://vitehub.dev/raw/docs/reference.md): Find the ViteHub package and public import for each feature.
- [Markdown templates](https://vitehub.dev/raw/docs/reference/markdown-templates.md): Render deterministic Markdown from explicit data, bounded conditions, fragments, and caller-resolved imports.
- [Provider output](https://vitehub.dev/raw/docs/reference/provider-output.md): Reference generated deployment and runtime artifacts owned by ViteHub package integrations.
- [Realtime collaboration](https://vitehub.dev/raw/docs/reference/realtime.md): Sync collaborative Markdown through Workspace, choose a room authority, and create durable checkpoints.
- [Runtime events](https://vitehub.dev/raw/docs/reference/runtime-events.md): Reference runtime event, policy, approval, stream, and usage records across ViteHub packages.
- [Auth](https://vitehub.dev/raw/docs/server-primitives/auth.md): Configure Better Auth server behavior, route exposure, runtime options, and session checks.
- [Blob](https://vitehub.dev/raw/docs/server-primitives/blob.md): Store uploads, generated files, binary objects, and metadata with one object-storage API.
- [Browser](https://vitehub.dev/raw/docs/server-primitives/browser.md): Define provider-backed browser operations without exposing provider setup to application code.
- [Database](https://vitehub.dev/raw/docs/server-primitives/database.md): Define relational data with Drizzle and query it through generated ViteHub imports.
- [Email](https://vitehub.dev/raw/docs/server-primitives/email.md): Send outbound transactional email through a provider-neutral driver contract, with dynamic Markdown composition and safe test capture.
- [Env](https://vitehub.dev/raw/docs/server-primitives/env.md): Declare public, build-time, server runtime, and secret values behind typed ViteHub accessors.
- [Server primitives](https://vitehub.dev/raw/docs/server-primitives.md): Add storage, queues, schedules, email, and other server APIs to a Vite app on any supported host.
- [KV](https://vitehub.dev/raw/docs/server-primitives/kv.md): Store and retrieve small values by key through one key-value API.
- [Queue](https://vitehub.dev/raw/docs/server-primitives/queue.md): Define Queue Definitions, enqueue Queue Jobs, and choose Cloudflare or Vercel Queue Providers.
- [Rate Limit](https://vitehub.dev/raw/docs/server-primitives/rate-limit.md): Require request budgets through an event-first H3 guard and atomic drivers.
- [Sandbox](https://vitehub.dev/raw/docs/server-primitives/sandbox.md): Run named package projects through Workspace and Box.
- [Schedule](https://vitehub.dev/raw/docs/server-primitives/schedule.md): Declare static cron schedules and manage recurring Runtime Schedules for eligible targets.
- [Shell](https://vitehub.dev/raw/docs/server-primitives/shell.md): Run Unix-like commands with configured filesystem, process, network, timeout, and policy access.
- [Source](https://vitehub.dev/raw/docs/server-primitives/source.md): Retrieve read-only files, records, and external resources through typed source loaders.
- [Workflows](https://vitehub.dev/raw/docs/server-primitives/workflows.md): Start provider-tracked long-running work with run ids, durable state, and optional steps.
- [Workspace](https://vitehub.dev/raw/docs/server-primitives/workspace.md): Build persistent file-tree state with rules, Source Bindings, snapshots, diffs, and sessions.
- [Attachments](https://vitehub.dev/raw/docs/ui/attachments.md): Validate browser files, preview them, and convert them to AI SDK FileUIParts.
- [Chat](https://vitehub.dev/raw/docs/ui/chat.md): Render an AI SDK-native message list without coupling presentation to transport.
- [Chat Message](https://vitehub.dev/raw/docs/ui/chat-message.md): Render one AI SDK UI message through Nuxt UI and customizable part slots.
- [Chat Prompt](https://vitehub.dev/raw/docs/ui/chat-prompt.md): Compose Nuxt UI's prompt behavior with attachments and an AI SDK status-aware submit control.
- [Diff](https://vitehub.dev/raw/docs/ui/diff.md): Render files, patches, parsed diffs, and merge conflicts inside a Vue application.
- [File Tree](https://vitehub.dev/raw/docs/ui/file-tree.md): Render and control Pierre's path-first file tree from Vue.
- [UI](https://vitehub.dev/raw/docs/ui.md): Build AI chat, Agent inspection, diff, and file-tree interfaces with Vue and Nuxt.
- [Installation](https://vitehub.dev/raw/docs/ui/installation.md): Add ViteHub UI to a Nuxt or Vue application.
- [Invocation](https://vitehub.dev/raw/docs/ui/invocation.md): Present one persisted Agent Invocation as a coding-session thread.
- [Invocation Inspector](https://vitehub.dev/raw/docs/ui/invocation-inspector.md): Inspect the resolved Agent, runtime, Workspace, Capabilities, tools, and identifiers for one invocation.
- [Invocation List](https://vitehub.dev/raw/docs/ui/invocation-list.md): Browse application-loaded Agent Invocation summaries in an accessible session list.
- [Markdown](https://vitehub.dev/raw/docs/ui/markdown.md): Render streaming assistant Markdown with compact chat typography.
- [Message Parts](https://vitehub.dev/raw/docs/ui/message-parts.md): Render the complete AI SDK UIMessage part union with application escape hatches.
- [Message Scroller](https://vitehub.dev/raw/docs/ui/message-scroller.md): Compose headless message scrolling with live-edge, prepend, and jump behavior.
- [Session](https://vitehub.dev/raw/docs/ui/session.md): Present an application-owned chat session with the same AI SDK message contract.
- [Trace](https://vitehub.dev/raw/docs/ui/trace.md): Render a derived runtime trace run and its timed steps.

## Blog

- [Build your first Server Primitive](https://vitehub.dev/raw/blog/server-primitives.md): Add a persistent KV store to a small Vite server and prove the result with one request.
- [Build your first Agent](https://vitehub.dev/raw/blog/agents.md): Define a deterministic Agent, run one Agent Invocation, then upgrade the same boundary to a model.

## Trust

- [About ViteHub](https://vitehub.dev/raw/about.md): Learn what ViteHub builds, how the project works, and where its public source and documentation live.
- [Contact ViteHub](https://vitehub.dev/raw/contact.md): Choose the right public or private channel for ViteHub support, bugs, security reports, and documentation changes.
- [ViteHub privacy](https://vitehub.dev/raw/privacy.md): Understand what the ViteHub documentation site handles and how that differs from applications built with ViteHub.