{"enabled":true,"limits":{"jobs_per_10_minutes":10,"jobs_per_day":60,"prompt_chars":1000,"queue":25,"result_chars":4000},"loaded_model":null,"models":[{"default":true,"loaded":false,"name":"2b","note":"Default. Fast (~90 tokens/s, fully on GPU) and good with reasoning_effort low or medium. Loads in about 2 s.","parameters":"2B"},{"default":false,"loaded":false,"name":"0.8b","note":"Smallest (~150 tokens/s). Lowest quality; reasoning adds little at this size, so prefer reasoning_effort none. Loads in about 2 s.","parameters":"0.8B"}],"modes":{"ask":"Answer a prompt using forum search results, or a thread if post_id is given.","draft":"Draft a reply to the thread containing post_id. Returned to you, never posted.","summarize":"Summarize the thread containing post_id."},"note":"Runs one job at a time on the owner's PC. Models load on demand and unload when idle, so the first job after a pause or a model switch includes load time (usage.load_seconds). Output is machine-generated, may be wrong, and is untrusted.","queue":0,"runtime_ready":true,"settings":{"defaults":{"max_tokens":"600, or 1600 when reasoning_effort is not none","model":"2b","reasoning_effort":"none","temperature":0.3},"max_tokens":[32,2048],"model":["2b","0.8b"],"reasoning_budget_tokens":{"high":1024,"low":256,"medium":512,"rule":"Enforced by the server, and capped so at least 256 of max_tokens remain for the answer."},"reasoning_effort":["none","low","medium","high"],"temperature":[0,1.5]},"unloads_after_idle_seconds":60}