Tools overview
FeatureQL tools live in four places: inside a query, as statements, in the Python package, and in wrappers you run from the demos repo. Pick the access that matches how you work.
| Access | What it is | How you get it |
|---|---|---|
| In the query | TYPEOF(), DESCRIBE(), RETURN TRACE, VERSION(), warning ACKs | FeatureQL — nothing extra |
| Statements | EXPLAIN, FORMAT, DESCRIBE, SHOW CREATE / DOCS / TESTS / FUNCTIONS | FeatureQL — nothing extra |
| Python library | BatchClient, Jupyter magic, AccessClient | pip install featuremesh |
| Demos repo | HTTP API server, MCP server, eval harness | Clone demos and adapt it to your local install |
pip install featuremesh does not start :8101 or :8100. BatchClient() talks to a local engine from the package. The HTTP API, MCP server, and eval runner live in the demos repo — you clone it, then wire ports, env, and backends to the FeatureMesh install you already run.
FeatureQL first
Stay on in-query tools and statement tools when you are already in a query. Use FORMAT and EXPLAIN from those pages when a tool needs more room.
Python from the package
from featuremesh import BatchClient
client = BatchClient()
client.query("RETURN F := 1;") # run and get a DataFrame
client.validate("RETURN F := 1;") # format, types, dry-run
client.help("getting_started") # concepts and examples
client.sltest(where="NAME LIKE '%array%#%'") # run conformance testsInstall and modes: Python library . Method detail: Python library tools .
Wrappers from the demos repo
If you cloned the demos stack and adapted it, those wrappers usually expose:
| Port | What it is for |
|---|---|
| 8101 | HTTP API — run FeatureQL end to end |
| 8100 | MCP — connect your AI client |
| 8001 | Registry alone — translation only, no query execution |
Ask the stack what it supports with GET http://localhost:8101/capabilities (host name may differ). Setup and routes: Demos repo tools .
There is no FeatureMesh-hosted MCP and no FeatureMesh-operated run of your warehouse SQL. Execution and MCP stay on machines you control. See FeatureQL with AI .
Follow-on how-tos
| Page | Install surface | When to read it |
|---|---|---|
| Conformance tests | Python sltest (HTTP/MCP only if the demos wrapper is running) | Run the 4,000+ documentation tests on your backends |
| Managing access | Python AccessClient | Orgs, projects, invitations, tokens |
| FeatureQL with AI | Python sandbox or local MCP | Connect an assistant, what to ask, what to review |
Related
- Language tour: FeatureQL for the Impatient
- Install: Python library
- AI setups: FeatureQL with AI