POST /v1/gb/remember | Extract a temporal knowledge graph from text and store it; or, with image, read an image and store the extraction | Remember, Artifacts, Temporal facts, Images, Chunking, Web search, Geo |
POST /v1/gb/timeseries, GET /v1/gb/timeseries, GET /v1/gb/timeseries/{id} | Store timeseries and anchor them to artifacts, entities and facts; list and read them. A remember request can also supply them, fetch them from the providers, or derive series_observation artifacts from them (timeseries) | Timeseries |
POST /v1/gb/query | Answer a question from the graph, with evidence | Query |
POST /v1/gb/datasets, GET /v1/gb/datasets?user_id=, DELETE /v1/gb/datasets/{name}?user_id= | Create, list and (soft) delete datasets | Datasets |
GET /v1/gb/datasets/{name}/graph?user_id= | Export a dataset’s whole graph as JSON | Graph export |
POST /v1/gb/ontologies, PUT /v1/gb/ontologies/{name} | Load an OWL ontology (RDF/XML), or update the one stored under a name (merge or replace) | Ontologies |
/v1/gb/mcp | MCP server (streamable HTTP) with remember and query as tools; off by default (mcp = true / serve --mcp), no authentication | MCP |
GET /health | Health check | |
GET /version | Version (and git commit) of the running binary | Versioning |
