Has anyone found a good way to use Clay to pass a Google Doc link or a link to a PDF document into Claygent or non-Claygent AI enrichment for analysis? It seems that Claygent is not able to open publicly accessibly document links and the non-Claygent AI enrichments don't have access to the web. I am trying to get an LLM to parse a document to pull out an array of information so that I can send it to another table for enrichment
Oops, something went wrong. Could you repeat your question?
I have tried to pass the PDF into Claygent as a Google Doc and as a public link to the doc in an S3 bucket
While Claygent is designed to access publicly available web data, it has limitations with directly accessing document links. Here's how you can work with documents in Clay: 1. For Google Docs: β’ Use our Google Docs integration to create or append content to docs β’ You can map your Clay data directly into Google Docs through the integration β’ Access this by clicking "Add enrichment" and searching for "Google Docs" 2. For document analysis: Instead of trying to pass document links to Claygent, consider these approaches: β’ Extract the relevant content from your documents and input it directly into your Clay table β’ Use webhooks to import document content into Clay for processing β’ Once the content is in Clay, you can then use our AI enrichments to analyze itRemember that Claygent is primarily designed to access and retrieve publicly available web data, but cannot access content behind paywalls or in protected documents, even if they're publicly shared. Sources: Google Docs integration | Documentation | Clay University Claygent: AI Web Scraper [Limitless Research] - Clay University
Is that what you were looking for?
Looks like you're still looking for help. If youβre dealing with a specific table, drop the URL below so we can help you quicker. Otherwise, someone from our team will be in touch soon!
I am running an automated process so trying to extract the information without have a human interact with it
how many files do you have? thinking out loud if you have openai api access
create an assitant for this task
upload files to it
have an http api column setup to call the assistant
send the data (in prompt) on the basis of which you'd like to retrieve info from the doc
other alt. i am thinking of is to use google app-scripts, but that's slightly tricky
Hey Ryan! Happy to help. I dug into this a bit more and confirmed Claygent can scrape a pdf file in your Google drive if you use the PDF link. I made a quick video showing an example of what PDF url should look like for this to work: https://www.loom.com/share/d28055ee928c4c9fb1150f3e40c25891 With that format, you can prompt Claygent to extract a list of the companies it finds or answer questions based on the content of the document. Let me know if that helps. https://downloads.intercomcdn.com/i/o/w28k1kwz/1377929572/6dbe4942274d99728de27062cfa0/CleanShot+2025-02-12+at+23_31_44%402x.png?expires=1739401200&signature=85bf15aed0a4c50fc5717d2e036c585c0cbcce7e418e34483621a438fcda77a9&req=dSMgEcB8lIRYW%2FMW1HO4zU8LasHiWeUEoxdlA%2B3QetP2ZyNvMSNPU%2Bz6k6dE%0AaJuc%0A
ooo that did work. so weird! i swear it was giving me errors before π€
π€·
i am using zapier to create the drive links and i do know there are like 3 different link options you can choose from and tbh i dont know the different and maybe claygent doesnt like some of them and as ive been playing around i could have been using differnet ones π
π Glad it worked! I made a follow up video here about outputting your results in JSON format: https://www.loom.com/share/c7458f1b481241ea9658c883d1df22cb Based on our tests, it seemed to consistently work for the exact url generated when you copy the "Copy link" in the "share" section. Does your embed link match that exactly?
the embed link must be that version, the other ones offered by zapier must be different
This thread was picked up by our in-app web widget and will no longer sync to Slack. If you are the original poster, you can continue this conversation by logging into https://app.clay.com and clicking "Support" in the sidebar. If you're not the original poster and require help from support, please post in 02 Support.