DevinVinson · Oct 6, 2026 · Oct 6, 2026 · Oct 6, 2026
diff --git a/docs.json b/docs.json
@@ -323,6 +323,7 @@
           {
             "group": "Release Notes",
             "pages": [
+              "openhands/usage/agent-canvas/release-notes/v1.25.0",
               "openhands/usage/agent-canvas/release-notes/v1.24.0",
               "openhands/usage/agent-canvas/release-notes/v1.23.0",
               "openhands/usage/agent-canvas/release-notes/v1.22.0",

diff --git a/openhands/usage/agent-canvas/agent-profiles.mdx b/openhands/usage/agent-canvas/agent-profiles.mdx
@@ -60,7 +60,7 @@

 ### Let the Agent Switch LLM Profiles

 The OpenHands profile editor includes a **"Let the agent switch LLM profiles"** toggle. When enabled, the agent is given the `SwitchLLMTool`, which lets it switch between available LLM profiles during a conversation. When disabled, the tool is removed from the agent's toolset.

 This toggle is version-gated: it appears only when the connected backend reports agent-server `1.31.0` or later. On older backends (for example, agent-server `1.29.0`–`1.30.x`) the toggle is hidden.

@@ -80,6 +80,25 @@
 
 See also [Model Context Protocol](/overview/model-context-protocol) for an overview of MCP support across OpenHands platforms.
 
+### Set a Profile's System Prompt
+
+The OpenHands profile editor includes a **System prompt** section. Choose one of three modes:
+
+- **OpenHands default** — Uses the built-in OpenHands instructions for coding, security, and tool use.
+- **Default + your instructions** — Adds your instructions after the built-in OpenHands instructions, which stay in place.
+- **Custom persona** — Replaces OpenHands' built-in persona and coding workflow. Memory, security and tool guidance, skills, repository instructions, and runtime context are still included.
+
+The default profile always uses the OpenHands system prompt. Create a named profile to customize it.
+
+### Choose a Profile's Tools
+
+A named OpenHands profile can choose its tools from the connected server's catalog:
+
+- **Standard tools** — The agent gets the tools the server considers standard. The list follows the server, so it can change when the server is upgraded.
+- **Choose tools** — The agent gets only the tools you select. Removing a tool removes the ability, not just the instruction to avoid it.
+
+The default profile always launches with the global tool set. Create a named profile to choose its tools. If the server's tool catalog cannot be loaded, the editor shows a retry action instead of the tool list.
+
 ## ACP Profiles
 
 Use an ACP profile when you want Agent Canvas to drive an external coding agent through the Agent Client Protocol.

diff --git a/openhands/usage/agent-canvas/canvas-extensions.mdx b/openhands/usage/agent-canvas/canvas-extensions.mdx
@@ -39,7 +39,7 @@

 <Tabs>
  <Tab title="Git Repository">
    1. Enter the Git source, such as `github:owner/repository`, or paste the app folder's browser URL from GitHub, GitLab, Bitbucket, Gitea, or Forgejo.
    2. Optionally enter a branch, tag, or commit in `Ref`.
    3. If the app is not at the repository root, enter its directory in `Repo path`.
    4. Select `Add app`.
@@ -75,6 +75,14 @@
 
 Only enable apps whose code and resolved revision you trust. An enabled app has the browser authority available to Agent Canvas and can use an authenticated helper to call the active Agent Server.
 
+### Update an App
+
+Apps installed from a Git repository or a backend-local path can be updated in place. Use the app's `Update` action to re-resolve its source and copy the new files over the installed version, without uninstalling and reinstalling it. This picks up changes even when the app's manifest version did not change. If the copy fails, the previously installed version keeps working.
+
+### App Icons
+
+An app can declare an `icon` in its manifest. When present, Agent Canvas loads the SVG from that path and shows it on the app's sidebar item and card.
+
 ## Build an App
 
 An app is a directory containing:
@@ -94,6 +102,7 @@
   "display_name": "Example dashboard",
   "version": "0.1.0",
   "description": "A project dashboard for Agent Canvas.",
+  "icon": "icon.svg",
   "entrypoint": "extension.js",
   "contributes": {
     "pages": [
@@ -110,7 +119,7 @@
 
 Use lowercase letters, numbers, and hyphens for app names and page IDs. Page paths must start with `/`, and every page ID and path must be unique within the app.
 
-The `entrypoint` must stay inside the app root. Bundle dependencies, CSS, and required assets into one browser ESM file; unresolved package imports and external runtime chunks cannot be loaded.
+The `entrypoint` must stay inside the app root. Bundle dependencies, CSS, and required assets into one browser ESM file; unresolved package imports and external runtime chunks cannot be loaded. If you declare an `icon`, it must also stay inside the app root; Agent Canvas serves it from `GET /api/canvas-extensions/installed/{name}/icon`.
 
 ### Register the Page
 

diff --git a/openhands/usage/agent-canvas/conversations.mdx b/openhands/usage/agent-canvas/conversations.mdx
@@ -7,7 +7,7 @@

 ## Child Conversations

 When an agent uses `launch_child_conversation`, Agent Canvas can launch a child conversation on a local or Cloud target. Local children can use either an isolated worktree or the parent's shared workspace. Cloud children use the repository and branch selected for the launch.

 The child remains linked to its parent, and its result is returned to the parent conversation. Agent Canvas validates the launch inputs before creating the child conversation.

@@ -15,7 +15,7 @@

 Use the conversation list controls to manage automation runs and visible tags:

 - Choose `All`, `Hide`, or `Only` to include, exclude, or show only automation-run conversations. You can further select individual automation names, including unnamed automations.
 - Pinned conversations remain visible when automation-run filtering would otherwise hide them.
 - Enable the `Tags` preference to show conversation tag chips. Tags are off by default. Each chip shows the tag key and value as a `Key: value` pair, such as `Artifacts: 1`. When there are more tags than fit, Agent Canvas shows a `+N` chip with the remaining count.

@@ -41,15 +41,33 @@

 ## Inline Markdown Artifact Previews

 When an agent creates a Markdown file, Agent Canvas renders it inline as a height-limited rich preview with an internal scrollbar instead of showing only the raw file content. Select `View` to open the full file in the Files drawer.

 ## Image Attachments
 
 When you upload an image as part of a message or an image is produced in a conversation, Agent Canvas displays it as a thumbnail. Click the thumbnail to open the image full size in a lightbox overlay. Dismiss the lightbox by pressing Escape, clicking the close button, or clicking the backdrop.
 
+## Voice Input
+
+Agent Canvas can add an optional microphone button to the chat input so you can dictate messages. Select the microphone to start dictation and select it again to stop; the transcript is added to the composer.
+
+By default, the microphone button uses your browser's speech recognition. To use your own transcription service instead, such as a local Whisper server, open `Settings > Application > Voice input` and enter an OpenAI-compatible endpoint:
+
+| Field | Purpose |
+|-------|---------|
+| `Base URL` | The transcription endpoint, for example `https://api.openai.com/v1` |
+| `API key` | The key for that endpoint |
+| `Model` | The transcription model; defaults to `whisper-1` |
+
+These voice input settings are stored in this browser only. The API key is kept in session storage so it is not persisted at rest.
+
+<Note>
+  If dictation fails, Agent Canvas reports that dictation failed and asks you to check microphone access and your voice input settings. Grant the browser microphone permission and confirm the endpoint details before retrying.
+</Note>
+
 ## Conversation Overview Panel
 
 The conversation overview panel displays project context for the active conversation, including workspace information, git state, and loaded resources such as skills, MCP servers, and automations.

 Toggle the overview using the info control in the conversation header. The panel peeks beside the chat area and closes when you open the Files drawer.

@@ -64,14 +82,28 @@

 ### Files View

 The **Files** tab is a focused file browser with open-file tabs and close controls. The file tree is resizable and persists its state across refreshes.

 Above the file tree, the active workspace path is displayed with a copy button. Hover the truncated path to see the full value in a tooltip, then click to copy it.

 <Note>
   The workspace path row is hidden when the conversation has no working directory.
 </Note>
 
+### File Discovery
+
+The Files panel discovers files per workspace. Open the file discovery settings to control what the tree includes:
+
+| Setting | Purpose |
+|---------|---------|
+| `Excluded directories` | One pattern per line. Names match anywhere; paths are relative to the workspace. Use `*/bin` to exclude nested `bin` directories while keeping the root `bin`. |
+| `Include symbolic links to files` | Include symlinks that point to files. |
+| `Maximum files (0 for unlimited)` | Cap the number of files the tree loads. Set to `0` to remove the cap. |
+
+Select `Restore defaults` to reset the settings. File discovery settings apply to this workspace on this backend, including other conversations that use the same workspace.
+
+When the number of discovered files exceeds the configured maximum, the tree is truncated and shows a finite limit notice: `Showing the first <count> files. This tree is incomplete. Adjust File discovery to see more.` Raise the maximum or narrow the exclusions to load more of the tree.
+
 ## Context Window Usage and Manual Compaction
 
 Agent Canvas shows a context-window meter in the composer that visualizes how much of the model's available context is in use. The meter fills as the conversation grows.
@@ -198,7 +230,7 @@

 ## Archive a Conversation

 Archiving a conversation hides it from the sidebar list without deleting it. The conversation's full history stays on the backend, and you can unarchive it at any time.

 **To archive a conversation:**

@@ -220,7 +252,7 @@
 </Note>

 <Note>
  After an installation switches from host-local conversation runtimes to Docker runtimes, conversations created under the previous local runtime have no Docker provisioning identity. Agent Canvas presents these conversations as archived. Their persisted event history remains readable, and Agent Canvas does not open a WebSocket connection for them.
 </Note>

 ## Related Guides

diff --git a/openhands/usage/agent-canvas/customize-and-settings.mdx b/openhands/usage/agent-canvas/customize-and-settings.mdx
@@ -44,7 +44,7 @@

 Server URLs and errors redact embedded secrets. Select **Update credentials** to edit a server; saving a corrected configuration refreshes its health status without reloading the page.

 On a cloud backend, health checks work for remote (SSE and Streamable HTTP) servers, which are tested through the app server. The Test button is not offered for stdio servers on a cloud backend, because those servers start inside the conversation sandbox rather than on the backend that serves the settings page. Local backends can test both remote and stdio servers.

 ## Settings

@@ -53,7 +53,7 @@
 | Section | Purpose |
 |---------|---------|
 | `Agent` | Agent Profile library and agent-specific capabilities |
-| `LLM` | Provider, model, API key, profile configuration, and provider connections |
+| `LLM` | Provider, model, API key, profile configuration, provider connections, and Model Router |
 | `Condenser` | Context compression and summarization behavior |
 | `Verification` | Approval, critic evaluation, and verification-related behavior |
 | `Application` | UI-level preferences and app behavior |

diff --git a/openhands/usage/agent-canvas/llm-profiles.mdx b/openhands/usage/agent-canvas/llm-profiles.mdx
@@ -56,7 +56,7 @@

 ### Local OpenAI-Compatible Endpoint

 A local server can be LM Studio, Ollama, vLLM, SGLang, or another service that exposes an OpenAI-compatible API. In the **Advanced** tab, enter the provider, exact model ID, endpoint base URL, and the required API key or a placeholder value when the server does not require one.

 The URL must be reachable from the **backend**, not only from your browser. For example, a backend in Docker cannot use `127.0.0.1` to reach a model server running on the host. Use the host address appropriate for that backend and confirm the endpoint's model inventory before saving.

@@ -66,7 +66,7 @@
 - **Base URL**: `http://host.docker.internal:1234/v1`
 - **API key**: `local-llm` or another placeholder value when the server does not require authentication

 See [Local LLMs](/openhands/usage/llms/local-llms) for LM Studio, Ollama, and other local-server examples.

 ### LiteLLM Proxy

@@ -100,6 +100,22 @@
 
 Edit a Provider Connection to rename it, rotate its API key, or change its base URL. The update applies to every linked profile. Before deleting a connection, re-link or change every profile that uses it; Agent Canvas prevents deleting a connection while profiles still reference it.
 
+### Bulk Add Models as Profiles
+
+When you want several models from one provider as saved profiles, use the bulk-add flow instead of creating each profile separately.
+
+1. Open `Settings > LLM` and select the add action to open `Add models as profiles`.
+2. Choose a provider from the searchable provider list.
+3. Select the models to add, or use `Select all`.
+4. Optionally select a `Provider connection` to link the new profiles to. Linked profiles inherit that connection's key; with `No connection (keyless)` selected, the profiles are created without credentials and need a key added later.
+5. Confirm to create the profiles.
+
+The same bulk-add action is available per provider connection. Use the connection's `Bulk add` action to add every selected model directly under that connection, so each new profile inherits the connection's credentials without repeating the key.
+
+<Note>
+Provider Connections are available on local agent-server backends only. On OpenHands Cloud, bulk-add creates keyless profiles unless you add a key to each one afterwards.
+</Note>
+
 ## Working with LLM Profiles
 
 LLM profiles are useful when you want different model setups for different tasks, such as:
@@ -135,7 +151,7 @@
 |---|---|---|
 | Provider is not recognized | Provider selection and model prefix | Use the matching configuration path above. |
 | Model format or identifier error | Exact model ID | Compare it with the provider or proxy model inventory. |
 | Local server cannot be reached | Base URL from the backend | Check host, port, and container or network reachability. |
 | Authentication or permission error | Key type and backend scope | Re-enter the key or follow the provider guide. |
 | Model cannot perform the task | Context and tool support | Choose a compatible model from the provider's recommendations. |


diff --git a/openhands/usage/agent-canvas/model-configuration.mdx b/openhands/usage/agent-canvas/model-configuration.mdx
@@ -37,6 +37,26 @@
 Provider Connections are available on local agent-server backends. The panel is hidden when using an OpenHands Cloud backend.
 </Note>
 
+## Model Router
+
+A **Model Router** is a profile that routes each task to the right LLM profile instead of pinning a conversation to a single model. Manage routers from the `Model Router` section of `Settings > LLM`.
+
+To create a router:
+
+1. Open `Settings > LLM` and select `Model Router`.
+2. Select `Add Model Router`.
+3. Choose a template: `Add OpenHands Router Pro`, `Add OpenHands Router Flash`, or `Custom`. The Custom option gives you a blank slate to build your own Model Router.
+4. Review the prompt template, model table, and then select one of your saved LLM connections so the router can add any missing LLM profiles needed for that router.
+5. Review the generated configuration, then save the router. Any needed LLM profiles will be added if you selected an LLM Connection which can take a few moments. 
+
+The Model Router settings include a **Run on first message** toggle. When enabled, the first message of every new conversation is routed through the active Model Router before the agent proceeds. The toggle is off by default but it will be enabled automatically when you add your first router.
+
+Model routing requires a backend that supports it. If the connected backend does not, the `Model Router` section reports that the backend does not support model routing yet and asks you to update the agent server.
+
+<Note>
+  Model Router is separate from the [LLM profiles](/openhands/usage/agent-canvas/llm-profiles) it routes to. Each router chooses among existing profiles rather than replacing them.
+</Note>
+
 ## Direct Provider or OpenHands Profile
 
 Use the `Basic` tab when you have an API key from Anthropic, OpenAI, Google, OpenHands, or another provider in the selector.
@@ -51,7 +71,7 @@

 ## Local OpenAI-Compatible Server

 Use the `Advanced` tab for LM Studio, Ollama, vLLM, SGLang, or another server that exposes an OpenAI-compatible API.

 1. Find the exact model ID served by your server, usually from its `GET /v1/models` endpoint.
 2. Enter `openai/<served-model-id>` as the model.

diff --git a/openhands/usage/agent-canvas/prebuilt-automations.mdx b/openhands/usage/agent-canvas/prebuilt-automations.mdx
@@ -3,9 +3,9 @@
 description: Get started quickly with a pre-built automation for common workflows.
 ---

 Agent Canvas ships with a set of pre-built automations for the most common agent workflows. Each one is a ready-to-use starting point — pick the one that fits your use case, connect it to the right backend, and you can have an automation running in minutes.

 ## Available Pre-built Automations

 | Automation | What It Does |
 |------------|-------------|
@@ -22,34 +22,34 @@

 In the `Automate` view, you can:

 - Browse existing automations
 - Inspect automation configuration and activity
 - Enable or disable automations
 - Edit an automation's LLM profile for future runs
 - Work with recommended automation flows

 ## How Creation Flows Usually Start

 The `Automations` view is mainly for browsing and managing automations that already exist.

 In practice, new automation setup starts in one of two ways:

 - From a conversation, where you ask OpenHands to `create an automation` for you
 - From a recommended automation flow in the `Automations` view

 For recommended automations that support a direct form setup, Agent Canvas checks the active backend's capabilities and any prerequisites, then guides you through the required input fields, a review step, and creation. If direct form setup is unavailable, it offers a conversation-assisted setup instead. Review the proposed configuration before creating an automation.
 
 Some catalog entries ship a **script bundle** — a packaged set of files that install as a deterministic automation — rather than a prompt-based preset. Script-bundle entries run their own logic for tasks like polling, deduplication, and fixed API calls, using the agent only for the parts that genuinely require judgment. When a catalog entry supports a bundle install, the setup form handles packaging and upload automatically; you just fill in the required fields.
 
-Catalog entries that accept repositories can also collect multiple repositories in a single field, so one automation can monitor several repos at once.
+Catalog entries that accept repositories can also collect multiple repositories in a single field, so one automation can monitor several repos at once. Automation templates appear on cloud backends as well as local backends. Where the built-in Git integration is offered, Agent Canvas presents it before the issue-to-PR templates.
 
 For a detailed walkthrough, see [Creating Automations](/openhands/usage/automations/creating-automations).
 
 <Note>
  Some recommended automations depend on integrations that cannot be auto-installed as MCP servers on this backend (for example, Jira's HTTP/OpenAPI-only integration). These appear on the recommendation card with a `Needs external setup` label. The `MCPs to connect` count only covers integrations the install flow can connect automatically. You must configure externally-hosted integrations yourself before the automation can use them.
 </Note>

 Automations run against the active backend. Use [Manage Backends](/openhands/usage/agent-canvas/backends) to see and switch which backend your automations run on.

 ## Edit an Automation's LLM Profile

@@ -59,7 +59,7 @@

 ## Select an Agent Profile for an Automation

 When the active backend advertises agent-profile support, the automation setup and edit dialogs include an optional **agent profile** selector. Choose a saved agent profile to use for automation runs instead of the default agent.

 When a saved agent profile is selected:


diff --git a/openhands/usage/agent-canvas/release-notes/v1.25.0.mdx b/openhands/usage/agent-canvas/release-notes/v1.25.0.mdx
@@ -0,0 +1,36 @@
+---
+title: Agent Canvas 1.25.0
+description: Release notes for Agent Canvas version 1.25.0
+---
+
+# Agent Canvas 1.25.0
+
+Released October 6, 2026.
+
+[View the full release on GitHub](https://github.com/OpenHands/OpenHands/releases/tag/v1.25.0).
+
+## Highlights
+
+- **Model Router settings** — Configure routers that route each task to the right LLM profile, with a **Run on first message** toggle that routes the first message of every new conversation. See [Configure a Model](/openhands/usage/agent-canvas/model-configuration#model-router).
+- **Bulk-add provider models as LLM profiles** — Create several LLM profiles at once from a provider picker, optionally binding them to a provider connection so they inherit its credentials. See [Manage LLM Profiles](/openhands/usage/agent-canvas/llm-profiles#bulk-add-models-as-profiles).
+- **Agent profile persona and instructions** — Give a profile its own system prompt, choosing between the OpenHands default, added instructions, or a custom persona. See [Agent Profiles](/openhands/usage/agent-canvas/agent-profiles#set-a-profiles-system-prompt).
+- **Voice input** — An optional microphone button in the chat input, backed by browser speech recognition or an OpenAI-compatible transcription endpoint. See [Conversations](/openhands/usage/agent-canvas/conversations#voice-input).
+- **Workspace file discovery** — Configure which files the Files panel discovers per workspace, including excluded directories, included symlinks, and a file-count limit. See [Conversations](/openhands/usage/agent-canvas/conversations#file-discovery).
+- **Update installed Apps** — Update an app installed from a local path or Git in place, without uninstalling and reinstalling it. See [Apps (Beta)](/openhands/usage/agent-canvas/canvas-extensions).
+- **Automations dashboard filter by creator** — Filter the Automations dashboard by anyone, you, or others. See [Managing automations](/openhands/usage/agent-canvas/managing-automations).
+
+## Fixes
+
+- Agent profiles pick their tools from the server's catalog, so the available tool list follows the connected server.
+- Automation templates are offered on cloud backends, and the built-in integration is offered before issue-to-PR templates.
+- Conversation title generation uses the running agent profile's LLM.
+- Focused settings pages keep their keyboard and viewport behavior on narrow screens.
+
+## Maintenance
+
+- Pinned agent-server and SDK to 1.53.0, with Extensions and Automation dependency bumps.
+
+## Full Changelog
+
+- [GitHub release notes](https://github.com/OpenHands/OpenHands/releases/tag/v1.25.0)
+- [Compare v1.24.0 to v1.25.0](https://github.com/OpenHands/OpenHands/compare/v1.24.0...v1.25.0)