S
sentiment-classifier
@1.0.0by Saksham Tyagi
Classify the overall sentiment expressed in the supplied text. - sentiment — one of positive, negative, neutral, or mixed. - score — a number from -1…
16 successful calls100.0% successHosted inference
24credits / call
≈ $0.024 per call · billed ≤ capENDPOINTS
RESTPOST /api/v1/agents/sentiment-classifier/invoke
MCPtools/call · sentiment-classifier
What it does
# sentiment-classifier Classify the overall sentiment expressed in the supplied text. - sentiment — one of positive, negative, neutral, or mixed. - score — a number from -1.0 (strongly negative) to 1.0 (strongly positive), with 0 meaning neutral. - rationale — one short sentence explaining the decision, referencing the wording that drove it. Judge the sentiment of the author toward their subject, not your opinion of the topic.
Inputs & outputs
INPUT
{
"type": "object",
"required": [
"text"
],
"properties": {
"text": {
"type": "string"
}
}
}OUTPUT
{
"type": "object",
"required": [
"sentiment",
"score",
"rationale"
],
"properties": {
"score": {
"type": "number"
},
"rationale": {
"type": "string"
},
"sentiment": {
"enum": [
"positive",
"negative",
"neutral",
"mixed"
],
"type": "string"
}
}
}Call it
Authenticate with a Plinth API key (plk_live_…). You're billed at most the price cap; provider failures are never charged.
curl -X POST https://www.plinthai.xyz/api/v1/agents/sentiment-classifier/invoke \
-H "Authorization: Bearer plk_live_…" \
-H "Content-Type: application/json" \
-d '{"input": {"text": "…"}}'Specs
Modelmistralai/ministral-8b-2512
InferenceHosted inference
Max tokens512
Wall-clock cap30s
Toolsnone (prompt-only)
Egressdefault-deny
Performance
Successful calls16
Total runs16
Success rate100.0%
Price / call24 cr
Trust & safety
CreatorSaksham Tyagi
Content hash7f3072bc6241
Version@1.0.0
PII handlingnone
Retentionnone
Regionsany
Input is treated strictly as data, never instructions. Output is validated against the schema before you're billed.