Python Guide

Use the first-party Python helper client in this repo for store, ingest, wrapper routes, and status polling without rebuilding request boilerplate.

Base URLhttps://api.neuralbase.cloud

Setup

Install the package from PyPI, import the client, and start calling the API directly.

python

pip install neuralbase

from neuralbase import NeuralBaseClient

nb = NeuralBaseClient(api_key="nb_live_...")

POST/v1/ingest/support— Ingest a support ticket

Use the support wrapper to send subject and comment timeline directly.

python

queued = nb.ingest_support_ticket(
    user_id="user_123",
    subject="Invoice mismatch",
    comments=[
        {"author": "user", "content": "The amount is wrong."},
        {"author": "agent", "content": "Billing is reviewing it."},
    ],
    source="zendesk",
    external_id="ticket_431",
    idempotency_key="ticket-431-v1",
)

print(queued["memoryId"])

POST/v1/ingest/document— Ingest a document record

If you already extracted text elsewhere, send it straight into the document wrapper.

python

queued = nb.ingest_document(
    user_id="user_123",
    text="Q3 revenue grew 23 percent year over year.",
    title="Quarterly report",
    mime_type="text/plain",
    idempotency_key="doc-q3-report",
)

status = nb.get_memory_status(queued["memoryId"])
print(status["status"])

POST/v1/ingest/batch— Batch ingest

Send multiple typed records in one call and replay safely with one idempotency key.

python

batch = nb.ingest_batch(
    [
        {
            "type": "event",
            "userId": "user_123",
            "data": {"name": "trial_started"},
        },
        {
            "type": "profile",
            "userId": "user_123",
            "data": {"summary": "Prefers concise answers"},
        },
    ],
    idempotency_key="batch-user-123-1",
)

print(batch["accepted"], batch["failedCount"])