fittings-text-sentiment

Original：🇺🇸 English

Translated

1 scripts

Score text from -1 (strongly negative) to +1 (strongly positive) with a label, handling negation, intensifiers and emphasis. Called as POST /v1/text/sentiment, it takes text and returns score, label, matches, positiveTerms. A support-triage or review-monitoring agent must decide whether a message is a complaint before it routes or escalates it, and a model call for one number costs more than this and is not reproducible. Reading this schema and dry-running the call are free and need no wallet; a real call costs $0.004, paid in USDC on Base over x402.

1installs

Sourcefittings-sh/fittings

Added on2026-08-17

NPX Install

npx skill4agent add fittings-sh/fittings fittings-text-sentiment

SKILL.md Content

View Translation Comparison →

Text sentiment score

text.sentiment

on

https://fittings.sh

. One endpoint, deterministic compute.

Score text from -1 (strongly negative) to +1 (strongly positive) with a label, handling negation, intensifiers and emphasis.

The call

POST https://fittings.sh/v1/text/sentiment

Field	Type	Required	Notes
`text`	string	yes	Text to score, up to 20000 characters

Example input:

json

{
  "text": "The delivery was late and the packaging was damaged, but support was very helpful."
}

Metered

The 402 quotes a ceiling — the price of the declared cap, 20000 characters — and settlement is for what the call actually used. A small request settles for less than it was quoted, so read the quote as an upper bound rather than a price.

Cost and provenance


Price	$0.004 per call, metered per characters
Source	none (pure compute)
Licence	none-needed

Why call it rather than answer from memory

A support-triage or review-monitoring agent must decide whether a message is a complaint before it routes or escalates it, and a model call for one number costs more than this and is not reproducible.

Calling it

A skill is loaded on its own, so this document has to be enough by itself. It does not assume the general

fittings

skill is also loaded.

call.mjs

ships alongside this file; if it is missing (this document reached you without the rest of the skill), fetch it once with

curl -sO https://fittings.sh/skill/call.mjs

.

Free, and needs no wallet. Read the schema and see what a call would cost:

bash

curl -s https://fittings.sh/catalog/text.sentiment | jq

Paying. The endpoint answers

402 Payment Required

with a signed quote, you sign an authorization, and a facilitator settles it — you never send a transaction and never pay gas. Before it spends anything, the client checks its own hash against the gateway and refuses rather than sign with logic the gateway has moved past:

bash

# Dry run. Prints the quote and pays nothing, because no key is set.
node call.mjs text.sentiment text="The delivery was late and the packaging was damaged, but support was very helpful."

# For real: a wallet holding USDC on Base, and nothing else.
export FITTINGS_PRIVATE_KEY=0x...
node call.mjs text.sentiment text="The delivery was late and the packaging was damaged, but support was very helpful."

Spending is never a side effect of curiosity: with no

FITTINGS_PRIVATE_KEY

the client fetches the quote and stops.

FITTINGS_MAX_PRICE

caps what it will pay without asking, and defaults to $0.05.

Use a throwaway wallet holding a few dollars, not a main one. The key sits in plaintext wherever the agent runs, and the whole cost of entry is USDC on Base — no account, no API key, and no ETH, because the payer signs off-chain and the facilitator broadcasts and pays the gas.

Writing your own client instead of using this one?

@x402/core

and

@x402/evm

on npm, and the

fittings

skill records the two things that cost real debugging time: the challenge is in a response header rather than the body, and it must be re-signed from the exact request being retried.

If this is not the right service,

https://fittings.sh/catalog/search?q=...

searches all of them for free.