Demandbase Python SDK Overview

Set the following environment variables before your application creates a Demandbase client:

• DEMANDBASE_CLIENT_ID (required)
• DEMANDBASE_CLIENT_SECRET (required)

Security best practices for Demandbase SDK Credentials:

• Do not commit DEMANDBASE_CLIENT_SECRET to source control.
• Store credentials in environment variables or a secret manager.

import demandbase

# Export credentials in your environment first
# export DEMANDBASE_CLIENT_ID=YOUR_ID
# export DEMANDBASE_CLIENT_SECRET=YOUR_SECRET

with demandbase.DBClient(timeout=60.0, retry_count=2) as client:
    try:
        # Example: list subscriptions (B2B API)
        subs = client.b2b_api.list_subscriptions(page=1, per_page=10)
        print(subs)
    except demandbase.DemandBaseAPIError as error:
        print(f"Demandbase API error: {error.http_status_code} - {error.error_message}")

The SDK returns a Pydantic SubscriptionList model, instead of a raw dict. Access model fields directly or serialize the model when you need JSON:

print(subs.subscriptions)
print(subs.model_dump())

Enable warning-level SDK logs:

import demandbase

demandbase.enable_logging()

Set a specific log level in code:

import demandbase

demandbase.enable_logging("INFO")

Supported log levels are DEBUG, INFO, WARNING, and ERROR.

By default, SDK logs are written to standard error (stderr). Pass a stream to send logs to another destination.

Write SDK logs to standard output:

import sys
import demandbase

demandbase.enable_logging("INFO", stream=sys.stdout)

Write SDK logs to a file:

import demandbase

log_file = open("demandbase-sdk.log", "a", encoding="utf-8")
demandbase.enable_logging("DEBUG", stream=log_file)

Set the log level before your application starts:

export DEMANDBASE_LOG_LEVEL=INFO

Configure logging once near application startup so the SDK uses the intended level and destination. UseDEBUGonly for troubleshooting because it can generate verbose request diagnostics. Authorization tokens are redacted from SDK log output.

Use with demandbase.DBClient() as client or instantiate and call close().

Call resource methods such as client.b2b_api.search_companies(...) with Pydantic request models.

Catch DemandBaseAPIError when an API request fails.

See Demandbase Python SDK Common Errors, Troubleshooting, and FAQs for error classes, handling patterns, and retry guidance.

Use client.data_import_api.submit_import_data(...) with CSV data as a pandas.DataFrame, raw bytes, or a local file path string. When you provide a path string, the SDK opens and reads the file.

List methods commonly accept a page number and a value that controls how many results are returned per page.

If a request receives a 401 because the access token is invalid or expired, DBClient refreshes the token and retries the request. The client also retries 429 and 5xx responses up to retry_count with exponential backoff. The initial delay is 30 seconds and doubles up to 600 seconds.

If retries are exhausted after a rate-limit response, the SDK raises DemandBaseAPIError. Check the error status and response headers to decide when to retry the request.