feat!: align media capabilities and add grounded-query tooling

.github/ISSUE_TEMPLATE/bug-report.yml

@@ -25,9 +25,9 @@ body:
     attributes:
       label: Affected capability
       options:
-        - Dialogue search
+        - Speech search
         - Scene search
-        - Action search (videoprism)
+        - Action and motion search
         - Actor matching
         - Multiple capabilities
         - Not capability-specific or unsure

.github/ISSUE_TEMPLATE/feature-request.yml

@@ -25,9 +25,9 @@ body:
     attributes:
       label: Related capability
       options:
-        - Dialogue search
+        - Speech search
         - Scene search
-        - Action search (videoprism)
+        - Action and motion search
         - Actor matching
         - Multiple capabilities
         - Not capability-specific or unsure

.github/release-intro.md

@@ -1,6 +1,6 @@
 ## Download VidXP
 
-{release_notice}VidXP turns video into searchable dialogue, sounds, scenes, actions, people, and
+{release_notice}VidXP turns video into searchable speech, sounds, scenes, actions, people, and
 inspectable evidence. Choose the desktop app for the guided local setup, or use
 the Python package and containers for command-line and server deployments.
 

.gitignore

@@ -2,6 +2,9 @@
 /venv
 /chroma_data
 /benchmark_runs/
+/benchmarks/codex-mcp/.promptfoo/
+/benchmarks/codex-mcp/results/
+/benchmarks/codex-mcp/workspace/
 /audio.wav
 /video.mp4
 /dist/

INSTALLATION_GUIDE.md

@@ -33,7 +33,7 @@ Approximate model downloads are:
 
 | Feature | Download |
 |---|---:|
-| Dialogue search | 2.63 GiB |
+| Speech search | 2.63 GiB |
 | Sound event search | 0.91 GiB |
 | Scene search | 1.43 GiB |
 | Action search | 0.93 GiB |
@@ -200,8 +200,8 @@ asking for confirmation. Download only selected features when preferred:
 
 ```bash
 vidxp prepare --modalities scene
-vidxp prepare --modalities dialogue,actor
-vidxp prepare --modalities videoprism  # action search
+vidxp prepare --modalities speech,actor
+vidxp prepare --modalities action
 vidxp prepare --modalities sound       # music and environmental sounds
 ```
 
@@ -232,14 +232,14 @@ vidxp index create <media-id>
 # Find a visual scene
 vidxp search scene "a yellow taxi on a city street"
 
-# Find an action or event (`videoprism` is the CLI name for action search)
-vidxp search videoprism "a person opens a door and walks outside"
+# Find an action or motion across multiple frames
+vidxp search action "a person opens a door and walks outside"
 
 # Find music or an environmental sound
 vidxp search sound "an alarm ringing"
 
 # Find something that was said
-vidxp search dialogue "the bread just came out of the oven"
+vidxp search speech "the bread just came out of the oven"
 ```
 
 Add `--media-id <media-id>` to a search command to restrict results to one
@@ -313,6 +313,30 @@ clients, or public access, use the supported server deployment instead.
 See [Local API and MCP server](docs/local-api.md) for authentication, uploads,
 and sharing behavior.
 
+## Optional local grounded answers
+
+VidXP search does not require a language model. To let CLI, HTTP, or MCP
+queries plan searches and draft grounded answers locally, enable **Local
+grounded answers** in VidXP Desktop setup. Desktop checks for a compatible
+loopback Ollama service, asks before installing Ollama through the supported
+Windows or macOS package manager when needed, and explicitly downloads the
+approved Qwen3.5 4B model. Linux setup links to Ollama's official installation
+instructions instead of running a privileged script.
+
+This optional feature follows Ollama's platform floor: Windows 10 22H2 or
+newer, or macOS 14 or newer. VidXP Desktop itself can still run without local
+grounded answers on older supported systems.
+
+The model is an additional approximately 3.4 GB download and has no per-run
+API charge or numbered hosted-model allowance. It uses local storage, memory,
+compute time, and electricity. Desktop configures the private service address
+for its browser, worker, API, Premiere, and generated MCP/Codex setup; there is
+no URL field to fill in. A command-line-only installation remains available
+for developers and custom deployments.
+
+The complete setup and its current evidence limitations are documented under
+[Enable local grounded answers](docs/local-api.md#enable-local-grounded-answers).
+
 ## Optional dependency extras
 
 Most users should choose one of the package profiles above. The individual
@@ -321,10 +345,10 @@ assembling a custom installation:
 
 | Extra | Adds |
 |---|---|
-| `dialogue` | Transcription, dialogue embeddings, and storage |
+| `speech` | Transcription, speech embeddings, and storage |
 | `sound` | Music and environmental-sound search and storage |
 | `scene` | Scene search and storage |
-| `videoprism` | Action search and storage |
+| `action` | Multi-frame action and motion search plus storage |
 | `actor` | Actor matching and storage |
 | `all` | Every built-in search feature |
 | `local-worker` | All search features plus local job processing |
@@ -404,6 +428,11 @@ to be indexed again. VidXP reports this instead of silently replacing a working
 index. Prepare the required models, re-index the affected videos, and keep the
 old repository until you have checked the replacement results.
 
+The current public capability names are `scene`, `action`, `sound`, `speech`,
+and `actor`. VidXP does not translate removed capability names.
+If an older repository reports an incompatible index schema, rebuild it using
+the current names.
+
 ## Data locations
 
 VidXP stores data outside the current working directory:

README.md

@@ -15,7 +15,7 @@
 </p>
 
 <p align="center">
-  <strong>Dialogue search · Sound search · Scene search · Action search · Actor grouping</strong>
+  <strong>Speech search · Sound search · Scene search · Action search · Actor grouping</strong>
 </p>
 
 <p align="center">
@@ -39,7 +39,7 @@
 
 VidXP makes one video—or an entire collection—searchable by meaning:
 
-- **Dialogue search:** type what you remember someone saying and jump to the
+- **Speech search:** type what you remember someone saying and jump to the
   matching moments.
 - **Scene search:** describe what appeared on screen and find the closest
   visual matches.
@@ -122,7 +122,7 @@ See the [Coolify guide](docs/deployment/coolify.md) for the complete setup.
 ## What you can do today
 
 - Build searchable libraries from individual videos or whole collections.
-- Find dialogue, sound events, visual scenes, and multi-frame actions by description.
+- Find speech, sound events, visual scenes, and multi-frame actions by description.
 - Ask grounded questions and inspect the supporting boards, frames, or clips.
 - Group recurring faces and render highlighted actor overlays.
 - Keep personal, client, or project libraries separate.
@@ -144,13 +144,13 @@ vidxp index create <media-id>
 vidxp search scene "a yellow taxi on a city street"
 
 # Find an action or event
-vidxp search videoprism "a person opens a door and walks outside"
+vidxp search action "a person opens a door and walks outside"
 
 # Find a sound event
 vidxp search sound "a dog barking over traffic noise"
 
 # Find something that was said
-vidxp search dialogue "the bread just came out of the oven"
+vidxp search speech "the bread just came out of the oven"
 ```
 
 Results include the source video, timestamps, match score, and the evidence
@@ -208,7 +208,7 @@ approximately 3 GiB.
 
 | Capability | Approximate model download |
 |---|---:|
-| Dialogue search | 2.64 GiB |
+| Speech search | 2.64 GiB |
 | Sound event search | 0.94 GiB |
 | Scene search | 1.43 GiB |
 | Action search | 0.93 GiB |

benchmarks/codex-mcp/.npmrc

@@ -0,0 +1 @@
+engine-strict=true