We are currently building something exciting and will be launching soon. Registration is temporarily closed. Launching soon — registration temporarily closed.

Text analysis API

Know what's in your text.

You send user messages, listings, or support text. Sentiqra returns structured signals — categories and labels with scores — for trust & safety, fraud, and support teams to route and review.

Marketplace messaging safety is one application. Explore the broader API catalog.

Inspect signals, not an opaque paragraph.

Sentiqra / Analysis

Sample

01 / Input text

“You are useless. I will find you and make you pay for wasting my time.”

02 / Signals & matched text

  • Threats & Intimidation

    “make you pay”

  • Insults & Derogation

    “You are useless”

Pinned config
sample-v1
Receipt ID
demo-0001

01 / Input text

“Pay outside the platform to secure the item. This offer expires in ten minutes.”

02 / Signals & matched text

  • Payment diversion

    “Pay outside the platform”

  • Social engineering

    “expires in ten minutes”

Pinned config
sample-v1
Receipt ID
demo-0001

Illustrative sample — not a live result. The API returns signals and verification metadata; it does not return matched-text spans.

01 / Built for review

Make the analysis part of the record.

Every result comes back with the configuration that produced it; paid plans add a receipt — a signed record of the result you can check later.

  1. 01 / Reproducibility

    Replay the result.

    Re-run the original input with the same pinned configuration (the fixed settings it ran under) and output profile to get the same result.

  2. 02 / Traceability

    Inspect the record.

    Inspect the signals and verify the attached receipt.

  3. 03 / Control

    Keep the decision.

    Apply your own rules, human review, and decision records.

What Sentiqra does not do

  • Make enforcement decisions — your policy and reviewers decide.
  • Provide calibrated probabilities — scores are signal-match strength.
  • Guarantee accuracy without testing on your own data.
  • Prove a result correct — verification proves a result is authentic, not that it is right.

02 / Put it to work

Find your workflow.

Featured workflow

Safer conversations. Reviewable decisions.

Bring repeatable text analysis into marketplace listings, messages, and moderation reviews.

Explore the marketplace workflow
  1. 01 Incoming message
  2. 02 Sentiqra signals + receipt
  3. 03 Your policy + human review

Security & fraud

Investigate suspicious messages.

Route suspicious communication into your existing investigation workflow.

Explore solutions

Customer intelligence

Understand customer conversations.

Identify documented customer-service and emotion signals in text.

Explore solutions
Browse the analyzer catalog
POST  /v1/analyze REST
{
  "text": "Your text to analyze",
  "analyzers": ["trust-safety"]
}
Structured signals + verification metadata Explore the response contract →

03 / Built to integrate

One API. A clear way forward.

Choose an analyzer, send text, and connect the returned signals to your application.

REST API · Documented response contracts

04 / Evaluate with confidence

Look closer. Then decide.

Review published evaluations, understand the limits, and test Sentiqra against your own workload. A receipt verifies an analysis record. It does not prove that a signal is correct.

The broader platform

One API. 12 generally available analyzers.

Trust & Safety is the starting point for this marketplace workflow. Explore other documented analyzers when your use case calls for them.

Your next step

See the signals. Inspect the evidence.

Start with a sample analysis, then explore how Sentiqra fits your workflow.