API Documentation

Overview

The repository now exposes a separate FastAPI service for creating verifications, polling report state, reading fact-check evidence, reading narrative analysis, and minting developer keys.

FeedAI uses a primary fact-check engine and multiple narrative lenses over the same evidence summary. If external engine keys are missing, the API falls back to local pipeline mode and records that assumption inside each report.

Endpoints

POST

/api/v1/verifications

Create a fact-check job with optional multi-model narrative analysis.

GET

/api/v1/verifications/:id

Read the full verification record, including factCheck, retrieval, and compatibility fields.

GET

/api/v1/verifications/:id/narratives

Read narrative views generated from the shared evidence summary.

GET

/api/v1/verifications/:id/sources

Read only fact-check sources and citations gathered by the retrieval orchestrator.

GET

/api/v1/api-keys

List API keys created by the current logged-in user.

POST

/api/v1/api-keys

Mint a developer API key.

DELETE

/api/v1/api-keys/:id

Delete one of the current user's API keys.

GET

/api/v1/usage

Read usage counters and engine availability.

Code Examples

Python

import requests

API_KEY = "feedai_your_api_key"

response = requests.post(
    "http://127.0.0.1:8000/api/v1/verifications",
    headers={"Authorization": f"Bearer {API_KEY}"},
    json={
        "query": "Did the IMF confirm that a country's GDP dropped by 10% in 2023?",
        "mode": "fact_check",
        "includeNarratives": True,
    },
)

job = response.json()
verification_id = job["verificationId"]

report = requests.get(
    f"http://127.0.0.1:8000/api/v1/verifications/{verification_id}",
    headers={"Authorization": f"Bearer {API_KEY}"},
).json()
print(report)

Environment Status

Signed in as Current user

Your API keys: 0

Your API Keys

No API keys created yet.

API Playground

Input

API Key