Data gaps is the result of integrating CrawlChat with your documentation.
People ask queries but you have not documented it in your docs, how do you come to know?
CrawlChat exactly flags those questions and let you know so that you can improve your documentation
It is actually nice to use MCPs.
Connected sentry MCP. Asked Claude to fix website errors. No need to copy/paste error messages, stack traces, etc. Errors fixed.
Connected @crawlchat. Asked Claude to find user's questions that were not documented. Documentation improved.
Integrate CrawlChat and find out what people are actually asking about your product, what do they exactly want.
These insights help you in improving your documentation and the product in general!
I am convinced that @crawlchat gonna be more and more developer and tech centric answering engine.
In that direction, I added an option to add a Github repo to the collection and let the CrawlChat plan stuff for you.
You can see in the image, it can also be used as a bot on Github PRs/issues. Or, your team can together plan out a task. It is very good and scans the code base deep!
I have tested this on large codebases as well and it does very good job.
I am sure this gonna be helpful for teams for mid size to large codebases.
I am a monorepo guy. I love it. When I started @crawlchat I didn't want to complicate my setup with workspaces and stuff.
I did a simple hack and had npm install shared package with just symlinks.
It works just fine so far! 11 months.
Yesterday, I needed something simple. Run lint on all apps & shared packages on GitHub actions.
As there was no common builder across apps, this was not so trivial. That's when I knew the time has arrived.
I knew this day would come but didn't want to complicate stuff from day 0. I explored Turborepo and this is exactly what I want.
A simple monorepo builder. Quickly got into migrating my repo to it. Mostly it was import path changes, package.json changes.
And the result? I could remove so many redundant files per app. Now it is all common. Everything from root folder. Neat and clean.
And yes, I could add the lint check just like that ;)
Love how @crawlchat categorises the questions and shows metrics on top of it!
Better view on the kind of questions people are asking and how well the docs are written!
@LoicBerthelot@crawlchat would be of great help here. It does more as well
- Answers repeated queries from docs
- 360 view of what people asking & insights
- Custom actions as you said to do specific actions
Here are the reasons Dmytro
- Any one can build anything these days with AI. If someone is paying for it, its for running it, maintaining it, evolving it continuously
- @crawlchat is for docs heavy software teams. It makes sense to open source it so that my distribution gets better
- Trust increases. Its b2b and built by indie. I guess making the code open increases the trust factor
- People who can pay will still pay. People who canโt pay will not pay anyways. In worst case they like it and refer it to others :)
We are 30 โญ๏ธ now :)