Activity
Mon
Wed
Fri
Sun
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
What is this?
Less
More

Memberships

Chase AI Community

27.5k members โ€ข Free

AI Automation Mastery

21.5k members โ€ข Free

AI Money Lab

38.6k members โ€ข Free

Brendan's AI Community

21.9k members โ€ข Free

๐ŸŽ™๏ธ Voice AI Bootcamp

7.8k members โ€ข Free

Tech Snack University

14.2k members โ€ข Free

AndyNoCode

25.6k members โ€ข Free

AI Automation Society

208.8k members โ€ข Free

1 contribution to AI Automation Society
Building & deploying AI Agents for other users
Hi, I'm very glad I came across this vibrant community of people developing AI Agents and automations. There are many examples of how one can build an AI Agent or even use multiple AI Agents to achieve great automations when they are deployed and used by the developer himself/herself. Specifically, when we try to automate Social Media posts and interactions, we use our own credentials and set up OAuth2 integration between those Social Media systems and our n8n workflow. But what about letting someone else use the deployed workflow with their own Social Media or other system credentials? Imagine you design a workflow to sell to users, the workflow can be deployed when the user completes a payment, and then the user should just authorise the workflow for Social Media (Twitter, LinkedIn, etc.) for the workflow to start posting and responding on the user's behalf. I think that OAuth2 authorization is also required by many services to allow the n8n workflow to perform activities on the user's behalf, for example sending Gmail messages, updating Google Docs Sheet etc. How to request the user to Authorize the workflow with their own credentials to allow the workflow to perform activities on the user's behalf when a workflow is deployed and can be used by any user that is granted access to it? Are there any examples of such workflows/templates/YT tutorials that you can share or refer to? Thank you very much.
0 likes โ€ข Feb 8
Thank you very much @Jessica Leighty for your response and explaining your approach. I must say this is a very non-conventional (and unexpected) approach that would definitely work, but I have some reservations: - it is not very scalable if you are going to have many users - gaining remote access to the end user's computer is a serious security concern - I would not expect many users to agree I think the process may be automated for each target service with their OAuth2 configuration by constructing a respective URL/link that the user will be asked to open, and which will send the user to a target system to grant access. In that scenario, the user performs the actions himself/herself. I think that what I described is the approach used in the LangGraph Agent Inbox - here is the part of the video that offers a user to authorise Twitter access: https://youtu.be/gF341XMN8cY?si=S_KJ3aU5iM0FKxAX&t=495
1-1 of 1
@leonid-shamis-3837
Software Architect interested in AI Agents and Agenting applications

Active 31d ago
Joined Feb 7, 2025
Powered by