Six intent-named tools grouped by audience. Each returns the standard MCPEnvelope.
Student wants to explore eligible universities before choosing where to apply.
| Inputs |
country,
region,
limit,
offset
|
|---|---|
| Returns | {universities: ShortlistUniversity[], total: int} |
| Backend | filter_universities + slice |
Inspect stored faculty profiles for one university before recommending contacts.
| Inputs |
university*,
limit,
offset
|
|---|---|
| Returns | {faculty: FacultyProfileV1[], total: int} |
| Backend | GET /faculty/{university} (paginated) |
Primary student intent — rank professors for a research field across the shortlist.
| Inputs |
research_field*,
country,
region,
limit
|
|---|---|
| Returns | RecommendationV1 (full schema in data) |
| Backend | POST /recommend |
Operator triggers an out-of-process scrape for one shortlist university.
| Inputs |
university*,
use_llm_fallback
|
|---|---|
| Returns | {run_id: string} |
| Backend | worker.run subprocess |
Poll a collection run until status is done or failed.
| Inputs |
run_id*
|
|---|---|
| Returns | {stage, status, profiles_found, profiles_stored, errors, universities_scraped} |
| Backend | runs collection |
Quick health probe before heavier tools.
| Inputs | |
|---|---|
| Returns | {api: ok, mongodb: ok|unavailable} |
| Backend | GET /health |
| MCP Tool | Backend | Method |
|---|---|---|
browse_universities |
filter_universities + slice | — |
view_faculty |
/faculty/{university} | GET |
recommend_professors |
/recommend | POST |
collect_faculty_data |
worker.run --run-id | subprocess |
check_collection_status |
runs collection | MongoDB |
check_status |
/health | GET |