nornweave-client 0.1.0

Python client SDK for NornWeave - Inbox-as-a-Service API for AI Agents

NornWeave Python Client

The NornWeave Python library provides convenient access to the NornWeave APIs from Python.

Table of Contents

• Installation
• Usage
• Async Client
• Exception Handling
• Advanced
  • Access Raw Response Data
  • Retries
  • Timeouts
  • Custom Client
• Pagination
• Releasing

Installation

pip install nornweave-client

Or install from source using uv (recommended) or pip:

cd clients/python

# Using uv (recommended)
uv pip install -e .

# Or with pip
pip install -e .

Usage

Instantiate and use the client with the following:

from nornweave_client import NornWeave

client = NornWeave(base_url="http://localhost:8000")

# Create an inbox
inbox = client.inboxes.create(name="Support", email_username="support")
print(f"Created inbox: {inbox.email_address}")

# List inboxes
for inbox in client.inboxes.list():
    print(inbox.name)

# Get an inbox
inbox = client.inboxes.get(inbox_id="...")

# Delete an inbox
client.inboxes.delete(inbox_id="...")

# Send a message
response = client.messages.send(
    inbox_id="...",
    to=["[e-mail]"],
    subject="Hello",
    body="This is the message body in **Markdown**."
)

# List messages
for message in client.messages.list(inbox_id="..."):
    print(message.content_clean)

# List threads
for thread in client.threads.list(inbox_id="..."):
    print(f"{thread.subject} - {thread.message_count} messages")

# Get thread with messages
thread = client.threads.get(thread_id="...")
for msg in thread.messages:
    print(f"{msg.role}: {msg.content}")

# Search messages
results = client.search.query(query="invoice", inbox_id="...")
for result in results:
    print(result.content_clean)

Async Client

The SDK also exports an async client so that you can make non-blocking calls to our API.

import asyncio
from nornweave_client import AsyncNornWeave

client = AsyncNornWeave(base_url="http://localhost:8000")

async def main() -> None:
    inbox = await client.inboxes.create(name="Support", email_username="support")
    print(f"Created inbox: {inbox.email_address}")

asyncio.run(main())

Exception Handling

When the API returns a non-success status code (4xx or 5xx response), a subclass of the following error will be thrown.

from nornweave_client import ApiError

try:
    client.inboxes.get(inbox_id="nonexistent")
except ApiError as e:
    print(e.status_code)  # 404
    print(e.body)         # {"detail": "Inbox not found"}

Specific exception types:

• NotFoundError - 404 responses
• ValidationError - 422 responses (invalid request data)
• RateLimitError - 429 responses
• ServerError - 5xx responses

Advanced

Access Raw Response Data

The SDK provides access to raw response data, including headers, through the .with_raw_response property.

response = client.inboxes.with_raw_response.create(
    name="Support",
    email_username="support"
)
print(response.headers)      # httpx.Headers
print(response.status_code)  # 201
print(response.data)         # Inbox object

Retries

The SDK is instrumented with automatic retries with exponential backoff. A request will be retried as long as the request is deemed retryable and the number of retry attempts has not grown larger than the configured retry limit (default: 2).

A request is deemed retryable when any of the following HTTP status codes is returned:

• 408 (Timeout)
• 429 (Too Many Requests)
• 5XX (Internal Server Errors)

Use the max_retries request option to configure this behavior.

client.inboxes.create(
    name="Support",
    email_username="support",
    request_options={"max_retries": 1}
)

Timeouts

The SDK defaults to a 60 second timeout. You can configure this with a timeout option at the client or request level.

from nornweave_client import NornWeave

# Client-level timeout
client = NornWeave(
    base_url="http://localhost:8000",
    timeout=20.0,
)

# Request-level timeout override
client.inboxes.create(
    name="Support",
    email_username="support",
    request_options={"timeout": 5.0}
)

Custom Client

You can override the httpx client to customize it for your use-case. Some common use-cases include support for proxies and transports.

import httpx
from nornweave_client import NornWeave

client = NornWeave(
    base_url="http://localhost:8000",
    httpx_client=httpx.Client(
        proxy="http://my.test.proxy.example.com",
        transport=httpx.HTTPTransport(local_address="0.0.0.0"),
    ),
)

For async:

import httpx
from nornweave_client import AsyncNornWeave

client = AsyncNornWeave(
    base_url="http://localhost:8000",
    httpx_client=httpx.AsyncClient(
        proxy="http://my.test.proxy.example.com",
    ),
)

Pagination

Paginated requests will return a SyncPager or AsyncPager, which can be used as generators for the underlying objects.

# Iterate over all items
for inbox in client.inboxes.list():
    print(inbox.name)

# Paginate page-by-page
for page in client.inboxes.list().iter_pages():
    print(f"Page with {len(page.items)} items")
    for inbox in page.items:
        print(inbox.name)

# Async iteration
async for inbox in async_client.inboxes.list():
    print(inbox.name)

Releasing

Instruction for Python client developers.

Version Bumping

cd clients/python

# Patch release (0.1.0 → 0.1.1)
uv version patch

# Minor release (0.1.1 → 0.2.0)
uv version minor

# Major release (0.2.0 → 1.0.0)
uv version major

Publishing

1. Bump version and commit:

   uv version patch
   git add pyproject.toml uv.lock
   git commit -m "chore: release v$(grep '^version' pyproject.toml | cut -d'"' -f2)"
   

2. Create and push tag:

   VERSION=$(grep '^version' pyproject.toml | cut -d'"' -f2)
   git tag "v$VERSION"
   git push origin main --tags
   

3. GitHub Actions automatically publishes to PyPI with attestations.

License

Apache-2.0