@artooi/ag-ui-web-component 0.2.0 → 0.2.2

package/CHANGELOG.md

@@ -7,6 +7,58 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.2.2] — 2026-06-02
+
+### Added
+- **Friendly tool-call card labels.** The built-in tools now carry `x-summary`
+  labels (`navigate_to_route` → "Navigate", `list_routes` → "List pages",
+  `read_page` → "Read the page", state-hook `read_*`/`set_*` → "Read/Update
+  <name>"). For tools whose schema never reaches the browser — **server-side
+  tools** (drf-mcp, `@tool` registry) — a new `toolSummaries: Record<string,
+  string>` property maps tool name → label as a fallback (e.g.
+  `chat.toolSummaries = { list_projects: "Search projects" }`).
+
+### Changed
+- A tool call that ends with **no client handler and no `TOOL_CALL_RESULT`** now
+  settles the card as **"No result returned."** instead of the misleading
+  "Executed on the server." (nothing executed it).
+
+### Fixed
+- **Incoming-text animations no longer double-fire.** Two distinct cases:
+  - *End of stream:* `data-text-animation="word"` wrapped the finished assistant
+    message into staggered `.word` spans on `TEXT_MESSAGE_END`, so a response
+    that had already streamed in re-animated itself one word at a time. The word
+    reveal now runs only when a message arrives **at once** (single text delta,
+    or an error bubble); a message streamed across multiple deltas keeps its
+    progressive reveal and isn't re-wrapped.
+  - *Reload from memory:* on rehydrate the whole transcript mounts at once, so
+    every restored assistant bubble animated its text in parallel (fade) or
+    re-wrapped word-by-word — wrong, since it's old content, not arriving. Restored
+    bubbles are now marked `message--restored`, excluded from the fade entrance
+    animation and never word-wrapped, so history appears statically.
+
+## [0.2.1] — 2026-06-02
+
+### Added
+- **Server-side tool results in the card.** The element now subscribes to
+  AG-UI's `TOOL_CALL_RESULT` event and settles the matching tool-call card with
+  the real server output (honouring the `data-tool-display` mode), instead of
+  the generic "Executed on the server." placeholder — which remains only as a
+  fallback when no result event is streamed.
+- **Tool calls and results survive a page refresh.** History replay now
+  reconstructs tool-call cards (from assistant `toolCalls`) and settles them
+  from the persisted `tool` result messages, so a rehydrated transcript shows
+  the full tool activity, not just the prose. Applies to every conversation
+  store (the data was already persisted; only the replay was incomplete).
+
+### Fixed
+- **Pending indicator could hang after a server-only round.** A round whose
+  tool calls were all server-side re-showed the "thinking" indicator after the
+  run had already finished, leaving it stuck. The indicator is no longer shown
+  speculatively for server tools, and a terminal `onSettled` guarantee clears
+  it (and re-enables input) on every run-loop exit — including the
+  `MAX_TOOL_ROUNDS` ceiling and errors.
+
 ## [0.2.0] — 2026-06-02
 
 ### Added
@@ -107,7 +159,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Notes
 - First release — exercising the automated npm OIDC publish pipeline end-to-end.
 
-[Unreleased]: https://github.com/Artui/ag-ui-web-component/compare/v0.2.0...HEAD
+[Unreleased]: https://github.com/Artui/ag-ui-web-component/compare/v0.2.2...HEAD
+[0.2.2]: https://github.com/Artui/ag-ui-web-component/compare/v0.2.1...v0.2.2
+[0.2.1]: https://github.com/Artui/ag-ui-web-component/compare/v0.2.0...v0.2.1
 [0.2.0]: https://github.com/Artui/ag-ui-web-component/compare/v0.1.1...v0.2.0
 [0.1.1]: https://github.com/Artui/ag-ui-web-component/compare/v0.1.0...v0.1.1
 [0.1.0]: https://github.com/Artui/ag-ui-web-component/releases/tag/v0.1.0

package/README.md

@@ -158,8 +158,14 @@ That's the whole integration: an `endpoint` attribute pointing at your AG-UI ser
 
 **Properties** (JS only, not attributes): `headers`, `autoConfirm`, `confirmPredicate`,
 `agentFactory`, `getTools`, `getContext`, `routeMap`, `navigate`, `getPageMap`,
-`autoInjectPageMap`, `conversationStore`, `navigationResult`, `skillContext`, plus the mirrors
-`endpoint` / `toolDisplay` / `collapsed`.
+`autoInjectPageMap`, `conversationStore`, `navigationResult`, `skillContext`, `toolSummaries`,
+plus the mirrors `endpoint` / `toolDisplay` / `collapsed`.
+
+`toolSummaries` is a `Record<string, string>` mapping tool name → a friendly card
+label, used when a tool has no `x-summary` in its own schema. Built-in and client tools
+should carry `x-summary` directly; this map is the seam for **server-side tools** (drf-mcp,
+the django-ag-ui `@tool` registry), whose schema never reaches the browser — e.g.
+`chat.toolSummaries = { list_projects: "Search projects" }`.
 
 **Methods**: `registerTool`, `registerStateHook`, `setSkills`, `appendMessage`, `newChat`,
 `setCollapsed`, `toggleCollapsed`.
@@ -185,8 +191,10 @@ an AG-UI `AbstractAgent`. On the first send the element builds a client (via the
 4. This repeats until the agent stops calling frontend tools, bounded by `MAX_TOOL_ROUNDS`.
 
 Tool calls the client doesn't own (server-side tools the server already executed) are left alone —
-the loop doesn't re-run them. The current tool catalog and context are read **fresh on every run**
-(`getTools()` / `getContext()`), so they always reflect the current page state.
+the loop doesn't re-run them, but their streamed `TOOL_CALL_RESULT` is rendered into the tool-call
+card (honouring `data-tool-display`), so server-side output is visible too. The current tool catalog
+and context are read **fresh on every run** (`getTools()` / `getContext()`), so they always reflect
+the current page state.
 
 ### Registering tools
 
@@ -406,7 +414,9 @@ history endpoint) for cross-tab/device durability:
 chat.conversationStore = new MyServerBackedStore();
 ```
 
-On mount the element rehydrates the transcript from the store, so the chat looks continuous.
+On mount the element rehydrates the transcript from the store, so the chat looks continuous —
+including tool-call cards and their results (reconstructed from the persisted `toolCalls` and `tool`
+messages), not just the text turns.
 
 **3. Resumable loop (`x-navigates` + `navigationResult`).** A tool whose schema carries
 `x-navigates: true` (use `X_NAVIGATES_KEY`; read back by [`isNavigates`](src/tools/is_navigates.ts))