Say why a call failed, and stop padding the ones that succeed

LLM_GUIDANCE.md

@@ -617,6 +617,8 @@ After showing the text tree, offer the user an interactive HTML version they can
 https://v3-cached.virtualflybrain.org/get_hierarchy_html?id=<ID>&relationship=<RELATIONSHIP>&direction=<DIRECTION>&max_depth=<DEPTH>
 ```
 
+The `get_hierarchy` tool itself omits that HTML by default — it is a second copy of the tree you already have in `descendants`/`ancestors`, and it is usually three quarters of the response. Pass `include_html: true` if you actually intend to render it.
+
 For example: `https://v3-cached.virtualflybrain.org/get_hierarchy_html?id=FBbt_00003686&relationship=subclass_of&direction=both&max_depth=2`
 
 The HTML page has a collapsible interactive tree with clickable links to VFB for every term.

README.md

@@ -273,7 +273,10 @@ docker run -p 3000:3000 virtualflybrain/vfb3-mcp:latest
 Retrieve detailed information about VFB terms using their IDs.
 
 **Parameters:**
-- `id` (string): VFB ID (e.g., "VFB_jrcv0i43")
+- `id` (string or array): One or more VFB IDs (e.g., "VFB_jrcv0i43"); an array is fetched in parallel and returned keyed by ID
+- `verbose` (boolean, optional): Return the raw response (default false)
+
+By default the response is trimmed, which roughly halves it. Each `Queries` entry keeps `query`, `label`, `preview_columns` and `output_format`; its argument schema (`takes`) and its `preview_results` block are dropped when that block has no rows, since it otherwise just repeats `preview_columns`. Each image entry keeps `id`, `label` and `thumbnail`; the other five file URLs are dropped, and the response says where they are — same directory as the thumbnail, named `thumbnailT.png`, `volume.nrrd`, `volume.wlz`, `volume_man.obj`, `volume.swc`. Nothing here is a guess about what you need: pass `verbose: true` and you get the untouched response.
 
 ### run_query
 Execute predefined queries on VFB data.
@@ -313,6 +316,18 @@ List the valid `facets_annotation` type names for the four type filters above, r
 
 If the deployed VFBquery predates the `/facets` endpoint, this falls back to a snapshot bundled with the server and says so — names absent from a snapshot result may still be valid.
 
+### get_hierarchy
+Traverse the ontology hierarchy for a VFB term.
+
+**Parameters:**
+- `id` (string): VFB term ID (e.g., "FBbt_00005801")
+- `relationship` (string): `part_of` for region/tissue structure, `subclass_of` for cell-type taxonomy
+- `direction` (string, optional): `descendants`, `ancestors`, or `both` (default `both`)
+- `max_depth` (number, optional): Levels to expand (default 1; `-1` for the full tree)
+- `include_html` (boolean, optional): Include the HTML rendering (default false)
+
+The endpoint returns the tree three times over: as data in `descendants`/`ancestors`, as a `display` string, and as an `html` document written for the VFB site's ROI browser — with `display_full` usually byte-identical to `display`. By default the `html` is omitted and `display_full` is dropped when it duplicates `display`, which takes a typical response from ~4.3 KB to ~1.4 KB. Pass `include_html: true` if you are embedding the site rendering.
+
 ### query_connectivity
 Query synaptic connectivity between neuron classes across all connectome datasets. At least one of `upstream_type` or `downstream_type` is required. Results are ranked strongest-first and paged: you get `limit` rows plus a `summary` computed over **every** connection found — weight min/max/total/mean, per-dataset counts, distinct neuron counts, and the top class pairs — so the totals stay true even though the rows are truncated. A broad query can find tens of thousands of connections, which is why paging is on by default.
 

TECHNICAL.md

@@ -186,9 +186,9 @@ it: one ranking, one place to fix it.
 ### MCP Tools Implementation
 
 #### get_term_info
-- **Input**: VFB ID string
-- **Output**: Term metadata, classifications, images, publications
-- **API Call**: POST to term info endpoint
+- **Input**: VFB ID string or array of IDs; optional `verbose`
+- **Output**: Term metadata, classifications, images, publications. Trimmed by default (empty query previews, query argument schemas and the non-thumbnail image file URLs are removed, and a `trimmed` field says what went and how to get it back) — about half the bytes. `verbose: true` returns the response untouched
+- **API Call**: GET `get_term_info`
 
 #### run_query
 - **Input**: VFB ID(s) and query type; optional `limit`/`offset` (paging) and `include_images`

package.json

@@ -1,12 +1,13 @@
 {
   "name": "vfb3-mcp",
-  "version": "1.10.0",
+  "version": "1.11.0",
   "description": "MCP server for VirtualFlyBrain API integration",
   "main": "index.js",
   "scripts": {
     "build": "tsc",
     "start": "node dist/index.js",
-    "dev": "tsc && node dist/index.js"
+    "dev": "tsc && node dist/index.js",
+    "test:live": "node test/live-smoke.js"
   },
   "repository": {
     "type": "git",

server.json

@@ -3,7 +3,7 @@
   "name": "org.virtualflybrain/vfb3-mcp",
   "title": "VirtualFlyBrain",
   "description": "MCP server for Drosophila neuroscience data from VirtualFlyBrain",
-  "version": "1.10.0",
+  "version": "1.11.0",
   "websiteUrl": "https://virtualflybrain.org",
   "repository": {
     "url": "https://github.com/Robbie1977/VFB3-MCP",