Skip to content
System One

sufianetaouil/every

Command line search that asks a yes/no question of every function in a codebase and ranks the answers by the probability that the answer is yes. It is grep whose pattern is a question, not embedding search, and the author measured 1,302 functions of gin-gonic/gin in 3.7 seconds for $0.018.

sufianetaouil/every on GitHub

Stars
7
Forks
0
Language
Python
License
MIT
Last push
17 Sep 2026

Read from GitHub on . Counts change daily.

How it uses Jev

except ImportError:
    sys.exit("Missing dependency. Run:  pip install httpx")

API_URL = os.environ.get("TYPESAFE_BASE_URL", "https://api.typesafe.ai").rstrip("/") + "/v1/systemone"
MODEL = os.environ.get("TYPESAFE_MODEL", "jev-latest")
USD_PER_INPUT_TOKEN = 0.042 / 1_000_000  # output tokens are free

WIDTH_LADDER = [1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048]
DEPTH_LADDER = [1, 2, 4, 8, 16, 32, 64, 96, 128]

# Prior token model, measured 2026-09-16 against jev-latest: input tokens per
# request ~= PRIOR_FIXED + PRIOR_VAR * width. Used only for the pre-flight cost
# estimate; refit from live data after phase 1 and reported at the end.
PRIOR_FIXED = 260
PRIOR_VAR = 165

View in reference/probe_jev.py

Open on GitHub

PrimitivesNoul
Use casesSemantic reranking

More like this

Use cases this is tagged with