query_arcanna_eventsLABSCO SUMMARY
An agent can operate an Arcanna deployment, not just read from it.
The coding loop is the unusual part: it writes Python, runs it, and saves the working block into the pipeline as an integration — alongside the ordinary work of querying events, driving jobs and sending decision feedback back to the model.
What it is
A management and query client for Arcanna's AI use cases. It reads processed events with filters, creates and controls jobs and integrations, feeds decision feedback back into the model, and generates, executes and saves Python as an Arcanna integration.
What you get
- Events processed by Arcanna, queried with multiple filters, plus the field list to build those filters from
- Jobs and integrations created, updated, retrieved and deleted
- The parameter schema an integration expects
- Jobs started, stopped and trained
- Feedback attached to an event's decision, so model accuracy moves on real corrections
- Python generated, executed, and saved into the pipeline as an integration
- A health check covering the server and the API key
Requirements
An Arcanna host and a management API key, both passed in the client config. Runs from the published Docker image or the PyPI package; Docker is the documented path.
Setup effort
One command plus a key — docker run -i --rm -e ARCANNA_MANAGEMENT_API_KEY -e ARCANNA_HOST arcanna/arcanna-mcp-server
USE CASES
01
Finding the events behind a decision you are questioning
Read which fields are filterable, then query the processed events down to the ones that matter.
02
Correcting the model when it got one wrong
Pull the event, then attach feedback to its decision so the correction lands where training will see it.
03
Getting a job running and keeping it that way
Read the jobs that exist, start the one you need, and stop it when ingestion should pause.
04
Retraining once enough feedback has landed
Trigger training on the job and check its state afterwards, rather than waiting on a scheduled run.
05
Standing up a new integration
Read the parameter schema the integration expects, then create or update it against that shape.
06
Turning a one-off script into part of the pipeline
Generate the code, run it to see it works, then save the block as an integration so it stops being a one-off.
14 tools
Listed in project docs. Your client may show a different list.
get_filter_fieldsupsert_resourcesget_resourcesdelete_resourcesintegration_parameters_schemagenerate_code_agentexecute_codesave_codestart_jobstop_jobtrain_jobadd_feedback_to_eventhealth_checkBEFORE YOU INSTALL
What to know first
Tools you get
14 tools named in the project's documentation.
Requires
- docker on PATH
- ARCANNA_MANAGEMENT_API_KEY in your environment
- ARCANNA_HOST in your environment
Where it runs
Both documented clients run the server on your machine.
INSTALLATION
Connect your client
Server setup
TERMINAL
docker run -i --rm -e ARCANNA_MANAGEMENT_API_KEY -e ARCANNA_HOST arcanna/arcanna-mcp-serverMCP Clients
Claude Desktop, Other MCP clients
JSON
{
"mcpServers": {
"arcanna-mcp-server": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"ARCANNA_MANAGEMENT_API_KEY",
"-e",
"ARCANNA_HOST",
"arcanna/arcanna-mcp-server"
],
"env": {
"ARCANNA_MANAGEMENT_API_KEY": "<YOUR_ARCANNA_MANAGEMENT_API_KEY>",
"ARCANNA_HOST": "<YOUR_ARCANNA_HOST>"
}
}
}
}
