syncteams-sdk 0.2.0

Python client for the SyncTeams Workflow API

SyncTeams Workflow SDK (Python)

A Python client for the SyncTeams Workflow API. Mirrors the capabilities of the JavaScript SDK, offering convenient helpers to execute workflows, monitor task status, and manage approval flows.

---

Installation

pip install syncteams-sdk

Requirements: Python 3.9 or newer

---

Quick Start

from syncteams_sdk import WorkflowClient

client = WorkflowClient(api_key="YOUR_API_KEY")

# Execute a workflow
result = client.execute_workflow(
    workflow_id="your_workflow_id",
    input={"email": "[e-mail]"},
    unique_id="customer-123",
)

task_id = result["taskId"]

# Wait for completion
final_status = client.wait_for_completion(
    task_id,
    poll_interval_ms=2_000,
    on_update=lambda status: print(f"Status: {status['status']}")
)

if final_status["status"] == "COMPLETED":
    print("Workflow completed successfully!")

---

Configuration

Option	Required	Default	Description
api_key	✅	–	Your SyncTeams API key
base_url	❌	https://develop.api.syncteams.studio	API base URL
timeout_ms	❌	30000	Request timeout in milliseconds
retry	❌	See below	Retry configuration for failed requests
default_headers	❌	{}	Extra headers merged into every request
user_agent_suffix	❌	–	Extra token appended to the default User-Agent

Retry Configuration

By default, the SDK retries transient failures with exponential backoff:

• Maximum attempts: 3
• Initial delay: 1 second
• Backoff factor: 2x
• Maximum delay: 30 seconds
• Retries on: 408, 425, 429, and all 5xx responses

You can override any subset of these values when constructing the client.

---

API overview

execute_workflow(workflow_id, input, unique_id=None)

Starts a workflow execution.

response = client.execute_workflow(
    workflow_id="your_workflow_id",
    input={"customer_id": "cust-123"},
)

print(response["taskId"], response["status"])

Returns the taskId and initial status.

get_task_status(task_id)

Fetches the latest status and the filtered event log for a task.

status = client.get_task_status(task_id)
print(status["status"], len(status.get("eventLogs", [])))

continue_task(task_id, decision, message=None)

Resumes a waiting workflow after an approval decision.

client.continue_task(task_id=task_id, decision="APPROVE")

client.continue_task(
    task_id=task_id,
    decision="REJECT",
    message="Missing documentation",
)

When decision is "REJECT", message is required.

wait_for_completion(task_id, *, poll_interval_ms=2000, max_wait_time_ms=600000, on_update=None, exit_on_waiting=False, terminal_statuses=None, stop_event=None)

Polls a task until it reaches a terminal status (COMPLETED, FAILED, or CANCELED).

final_status = client.wait_for_completion(
    task_id,
    poll_interval_ms=1000,
    on_update=lambda payload: print("Status:", payload["status"]),
)

execute_and_wait(workflow_id, input, **options)

Convenience method that starts a workflow and optionally handles approvals via on_waiting.

def handle_waiting(status):
    # Perform approval logic
    client.continue_task(task_id=status["taskId"], decision="APPROVE")
    return True

result = client.execute_and_wait(
    workflow_id="wf-123",
    input={"amount": 500},
    on_waiting=handle_waiting,
)

If on_waiting returns False, polling stops and the SDK returns the current status (even if still waiting).

---

Error Handling

The SDK raises WorkflowAPIError for API failures. It exposes the HTTP status, headers, response payload, and request metadata to simplify debugging.

from syncteams_sdk import WorkflowAPIError

try:
    client.execute_workflow(workflow_id="invalid", input={})
except WorkflowAPIError as error:
    print("API error:", error.status, error.data)

Transient errors (timeouts, rate limits, server errors) are automatically retried according to the configured policy.

---

Webhooks

You can receive workflow updates via webhooks instead of polling:

from flask import Flask, request
from syncteams_sdk import WebhookEventPayload

app = Flask(__name__)

@app.post("/webhooks/syncteams")
def handle_webhook():
    payload: WebhookEventPayload = request.get_json(force=True)
    print("Task", payload["taskId"], "status:", payload["status"])
    return ("", 200)

---

Development

Install dependencies and run tests:

pip install -e .[dev]
pytest

---

License

MIT