langflow/docs/docs/deployment/backend-only.md

Backend-only

You can run Langflow in --backend-only mode to expose your Langflow app as an API, without running the frontend UI.

Start langflow in backend-only mode with python3 -m langflow run --backend-only.

The terminal prints Welcome to ⛓ Langflow, and a blank window opens at http://127.0.0.1:7864/all. Langflow will now serve requests to its API without the frontend running.

Prerequisites

• Langflow installed

• OpenAI API key

• A Langflow flow created

Download your flow's curl call

1. Click API.
2. Click curl > Copy code and save the code to your local machine. It will look something like this:

curl -X POST \
    "http://127.0.0.1:7864/api/v1/run/ef7e0554-69e5-4e3e-ab29-ee83bcd8d9ef?stream=false" \
    -H 'Content-Type: application/json'\
    -d '{"input_value": "message",
    "output_type": "chat",
    "input_type": "chat",
    "tweaks": {
  "Prompt-kvo86": {},
  "OpenAIModel-MilkD": {},
  "ChatOutput-ktwdw": {},
  "ChatInput-xXC4F": {}
}}'

Note the flow ID of ef7e0554-69e5-4e3e-ab29-ee83bcd8d9ef. You can find this ID in the UI as well to ensure you're querying the right flow.

Start Langflow in backend-only mode

1. Stop Langflow with Ctrl+C.
2. Start langflow in backend-only mode with python3 -m langflow run --backend-only. The terminal prints Welcome to ⛓ Langflow, and a blank window opens at http://127.0.0.1:7864/all. Langflow will now serve requests to its API.
3. Run the curl code you copied from the UI. You should get a result like this:

{"session_id":"ef7e0554-69e5-4e3e-ab29-ee83bcd8d9ef:bf81d898868ac87e1b4edbd96c131c5dee801ea2971122cc91352d144a45b880","outputs":[{"inputs":{"input_value":"hi, are you there?"},"outputs":[{"results":{"result":"Arrr, ahoy matey! Aye, I be here. What be ye needin', me hearty?"},"artifacts":{"message":"Arrr, ahoy matey! Aye, I be here. What be ye needin', me hearty?","sender":"Machine","sender_name":"AI"},"messages":[{"message":"Arrr, ahoy matey! Aye, I be here. What be ye needin', me hearty?","sender":"Machine","sender_name":"AI","component_id":"ChatOutput-ktwdw"}],"component_display_name":"Chat Output","component_id":"ChatOutput-ktwdw","used_frozen_result":false}]}]}%

Again, note that the flow ID matches. Langflow is receiving your POST request, running the flow, and returning the result, all without running the frontend. Cool!

Download your flow's Python API call

Instead of using curl, you can download your flow as a Python API call instead.

1. Click API.
2. Click Python API > Copy code and save the code to your local machine. The code will look something like this:

import requests
from typing import Optional

BASE_API_URL = "http://127.0.0.1:7864/api/v1/run"
FLOW_ID = "ef7e0554-69e5-4e3e-ab29-ee83bcd8d9ef"
# You can tweak the flow by adding a tweaks dictionary
# e.g {"OpenAI-XXXXX": {"model_name": "gpt-4"}}

def run_flow(message: str,
  flow_id: str,
  output_type: str = "chat",
  input_type: str = "chat",
  tweaks: Optional[dict] = None,
  api_key: Optional[str] = None) -> dict:
    """
    Run a flow with a given message and optional tweaks.

    :param message: The message to send to the flow
    :param flow_id: The ID of the flow to run
    :param tweaks: Optional tweaks to customize the flow
    :return: The JSON response from the flow
    """
    api_url = f"{BASE_API_URL}/{flow_id}"

    payload = {
        "input_value": message,
        "output_type": output_type,
        "input_type": input_type,
    }
    headers = None
    if tweaks:
        payload["tweaks"] = tweaks
    if api_key:
        headers = {"x-api-key": api_key}
    response = requests.post(api_url, json=payload, headers=headers)
    return response.json()

# Setup any tweaks you want to apply to the flow
message = "message"

print(run_flow(message=message, flow_id=FLOW_ID))

3. Run your Python app:

python3 app.py

The result is similar to the curl call:

{'session_id': 'ef7e0554-69e5-4e3e-ab29-ee83bcd8d9ef:bf81d898868ac87e1b4edbd96c131c5dee801ea2971122cc91352d144a45b880', 'outputs': [{'inputs': {'input_value': 'message'}, 'outputs': [{'results': {'result': "Arrr matey! What be yer message for this ol' pirate? Speak up or walk the plank!"}, 'artifacts': {'message': "Arrr matey! What be yer message for this ol' pirate? Speak up or walk the plank!", 'sender': 'Machine', 'sender_name': 'AI'}, 'messages': [{'message': "Arrr matey! What be yer message for this ol' pirate? Speak up or walk the plank!", 'sender': 'Machine', 'sender_name': 'AI', 'component_id': 'ChatOutput-ktwdw'}], 'component_display_name': 'Chat Output', 'component_id': 'ChatOutput-ktwdw', 'used_frozen_result': False}]}]}

Your Python app POSTs to your Langflow server, and the server runs the flow and returns the result.

See API for more ways to interact with your headless Langflow server.