The Pulp API measures persuasion. Send it language, a pitch, a memo, a transcript, an ad, anything people say to move other people, and it returns a structured measurement of persuasive force: a single force reading, the eight dimensions behind it, and a diagnosis that reads like an edit.
You will go from nothing installed to your first score in about a minute. Everything below is Python.
The Python SDK installs from PyPI. One line and you are ready to score.
pip install pulp-sdkEvery request carries a bearer key. Keys are scoped per environment: pk_test_ keys hit the sandbox, pk_live_ keys hit production. Store your key as an environment variable rather than pasting it into code, so it never lands in a commit.
export PULP_API_KEY="pk_live_..."The SDK reads PULP_API_KEY from the environment automatically, so you never pass the key in code.
The canonical first call: hold a message up to the light and read its force.
from pulp import Pulp
client = Pulp() # reads PULP_API_KEY from the environment
reading = client.radiance.rootcause.score(
text="Our proposal saves the district four million dollars "
"without closing a single school."
)
print(reading.force) # 74
print(reading.dimensions.logos) # 88
print(reading.dimensions.pathos) # 52That is the whole thesis in three fields: the persuasive force of the message (reading.force) and the dimensions that produced it (reading.dimensions.logos, reading.dimensions.pathos, and the six others). Ask for reading.diagnosis and you get not a grade but an edit: where the message is weak, what kind of weakness it is, and which moves would repair it.
That is it. You just did, in one function call, what your organization currently does with a conference room, four opinions, and an hour nobody gets back.
Scoring is the center, but a full persuasion workflow is a loop, and it closes inside one API:
Next: the Server-Side SDK reference, which documents every package in full.