Filip Michalsky

https://huggingface.co/docs/diffusers/main/en/api/pipelines/audioldm2 Seems not to support TTS, is it in plan to add your two tts checkpoints? thank you

The extract() method stops executing when encountering the first piece of relevant information. This is not ideal since the entire information we seek can be across multiple chunks. this is...

create a searchable vector DB of all elements vs just chunking

2

It's sometimes hard to find the right elements when navigating the website in chunks - things can reload etc as scroll down. What if we have a hashing/embedding algorithm for...

Allow Developers to change DOM parsing method

1

As a user of Stagehand, I might want to change the method of DOM parsing: from Travis: * Have some modularity (backstage) that allows developers to choose which method they...

Creating CONTRIBUTING.md doc

1

we need better instructions of the workflow to submit the code from outside contributors Langchain has a great doc: https://github.com/langchain-ai/langchainjs/blob/main/CONTRIBUTING.md might be an overkill but we need more than what...

this should come as stagehand expands

better manage types from LLM return calls

1

hey stagehand team, seeing a few methods return an any type -- this isn't good practice; esp in like the LLM client where you return instead of just a string...

running 10+ evals on different browsers in parallel locally can become computationally challenging. Braintrust offers a concurrency semaphor: https://www.braintrust.dev/docs/guides/evals/run ``` import { Factuality, Levenshtein } from "autoevals"; import { Eval...

Sometimes Stagehand continues with the next action even thought we have waitfordomsettle() function.