1. Grab your partner token
Tokens are scoped per organisation. Keep them secret—treat as you would any bearer credential.
API Manual
Integrate Greenophy’s ESG substantiveness classifier into your workflows. Authenticate with partner tokens, send paragraphs for evaluation, and receive structured labels for clarity and compliance.
Quickstart
Authenticate, send text, receive labels
Follow the steps below to call the hosted substantiveness model. Helpers in Python are available, but the API works with any HTTPS client.
2. POST to /api/substantiveness
Send paragraphs or multi-sentence disclosures in text. The server performs segmentation and returns JSON labels.
3. Store the response
Results contain structured labels, confidence scores, and supporting detail. Persist the request_id for traceability.
Authentication
Use your partner token in the
Use your partner token in the X-API-Key header
All requests require HTTPS with a partner token. Rated requests over limit return 429; invalid token returns 401.
Required headers
Content-Type: application/jsonX-API-Key: <partner-token>
Base URL
https://greenophy-service-70959934638.europe-west1.run.app
All endpoints below should be appended to this base path.
Endpoints
Available classifiers and payload contract
Both classifiers accept JSON payloads with the same schema. The substantiveness model returns ESG-focused labels, while the generic variant provides baseline categories.
POST
/api/substantiveness
Description ESG substantiveness classifier used by supervisory teams. Highlights evidence-backed statements and flags vague language.
Returns
Concrete Actions
Demonstrated Impact
Time-bound Commitments
Verification & Oversight
Symbolic
POST
/api/generic_classification
Description Neutral generic classifier for benchmarking or internal baselines where ESG-specific nuance is unnecessary.
Returns
Concrete Actions
Demonstrated Impact
Time-bound Commitments
Verification & Oversight
Symbolic
Payload
Required body fields
Send a JSON object with disclosure text and optional metadata. The API trims whitespace and rejects empty inputs with 400.
POST body schema
- text (string) – Paragraph or multi-sentence disclosure. Required.
- request_id (string, optional) – Echoed in response metadata for reconciliation.
- source (string, optional) – Tag upstream system or workflow (“dashboard-upload”, “csrd-batch”).
Response structure
- labels – Array of label objects with name, confidence, and supporting span.
- meta – Includes
request_id, total tokens, processing time. - version – Semantic version of the deployed model.
Examples
Sample request and response
Use the Python client below to call the hosted service. You can also issue direct HTTPS POST requests if you prefer.
Request
# pip install greenophy
from greenophy import SubstantivenessClient
# Hosted client defaults to the managed Greenophy endpoint
client = SubstantivenessClient(
api_key="YOUR_API_KEY_HERE", # Replace with your actual API key
timeout=120,
)
response = client.classify_esg_text(
"We transitioned 70% of our fleet to EVs in 2023."
)
for item in response["results"]:
print(f"{item['label_name']}: {item['sentence']}")
print("Quota remaining:", response["meta"]["quota_remaining"])
# Swap to classify_generic_text(...) for the neutral model
Response
{
"results": [
{
"index": 0,
"sentence": "In 2023 we transitioned 70% of our fleet to EVs.",
"label": 1,
"label_name": "Concrete Actions"
}
],
"meta": {
"count": 1,
"processing_seconds": 0.41,
"source": "text",
"quota_remaining": 998
}
}
Operations
Support and reliability
We monitor uptime and classification drift continuously. Reach out for access upgrades, incident response, or detailed audit logs.