Skip to content

Params

GET /v1/params is the live catalogue of everything you can currently vary on a completion: which modes are available (and whether each accepts client tools), which models you may select, every per-call setting with its bounds, and the hard limits. Consult it rather than assuming what is enabled — availability changes without notice, and this read is free.

It is composed from the same constants the validators enforce, so it never drifts from what the API actually accepts. If you skip it, the API's 422 responses still name the accepted values.

GET /v1/params

Return your live view of modes, models, settings and limits.

curl https://api.humx.ai/v1/params \
  -H "Authorization: ApiKey $COMPANIONS_API_KEY"

import os, requests

r = requests.get(
    "https://api.humx.ai/v1/params",
    headers={"Authorization": f"ApiKey {os.environ['COMPANIONS_API_KEY']}"},
)
r.raise_for_status()
data = r.json()
for m in data["modes"]:
    print(m["mode"], "enabled" if m["enabled"] else "disabled")

Response

200

Field Type Description
modes array Each {mode, summary, enabled, accepts_tools} — the consultable modes with one-line explanations and live availability.
models object Byte-for-byte the GET /v1/models payload: slugs, sampling ranges, the default.
settings object Every per-call settings.* knob (plus use_turns), each with its type, bounds and a one-line description.
limits object The named caps: max_participants, max_crumb_depth, max_declared_tools, max_prompt_chars, max_history_chars, max_messages. 
{
  "modes": [
    {"mode": "answer", "summary": "One companion answers your question in its own voice...", "enabled": true, "accepts_tools": true},
    {"mode": "parallel", "summary": "Each chosen companion answers the same prompt independently...", "enabled": true, "accepts_tools": true}
  ],
  "models": {"default_model": "...", "models": [{"slug": "...", "temperature_min": 0.0, "temperature_max": 2.0}]},
  "settings": {"crumb_depth": {"type": "integer", "minimum": 1, "maximum": 5, "description": "..."}},
  "limits": {"max_participants": 15, "max_crumb_depth": 5, "max_declared_tools": 32}
}