Introducing OpenLogs. A key part of my local dev stack. Stop copy & pasting your terminal logs into an agent. Now just prefix your command with ol and your agent can see everything. Log visibility was a huge bottleneck for local development - agents were unaware when something isn't... working. Now they can just check the logs themselves and fix - no more human involvement. Give it a try and let me know any feedback!show more

We did something similar in @deeptracer_dev. Kudos to providing dev-tools that are safer for AI usage! Our solution to models not understanding logs were more simpler: - creating MCPs - AI Investigation in the platform take a look at:

@deeptracer_dev Looks fire!

Curious how this is different from using an agent within Cursor where it also has direct access to the terminal?

I assume it solves a similar problem if cursor can read the terminal logs directly - right now opencode, Claude code, codex etc don’t have that solution

Yup, it’s what I really like about Cursor, I can use Claude and Codex models, but they can see the terminal logs, saves a lot of back and forth

Now you can use all the agents with terminal logs haha

I do like the sound of that

Lmk how u get on!

The biggest productivity sink in AI-assisted dev isn't the model — it's the context handoffs. Copy-pasting logs, switching windows, reformatting errors manually. Every friction point between what happened and what the agent sees is where useful signal gets dropped. Tools that close this gap are where the real leverage is.

🔥🔥

What I want is that wallpaper please

You can get from an app called backdrop!

Straight up awesome job dude... Even flashing that animated background

Thank u bro!!

@SamNewby_ 🔥🔥

@SamNewby_ 🤝🤝🤝

ngl log visibility is one of those things you don't realize is slowing you down until someone solves it. been manually piping output to files just so the agent can read them. does ol handle streaming output or does it wait for the command to complete?

It streams it!

interesting

Thanks!

log visibility is genuinely one of the most annoying parts of working w/ agents rn. half my debugging time is reconstructing what the agent actually did. gonna try this

Cheers! Lmk how u get on!

simple idea, real pain point. i've been piping logs through a named pipe into a file the agent watches but the `ol` prefix is cleaner. does it buffer or stream? wondering how it handles long-running processes where the agent needs to react mid-run rather than after.

Should all work fine - it streams the logs into a file as they come in - lmk if u have any issues!

okay super cool thanks 😊

looks super useful DX via the trail command is dope for agents smart way to get real-time insights let's launch you on anytime that helps mate

Wow, really nice. Congrats on the launch bro.

Thanks bro!!

fire, are you accepting donations via GitHub by any chance?

I can set it up if you wanna support!! Appreciated 😁

for sure, send me your github link!

Just set this up! - thanks man❤️‍🔥

looks really good!

Thank you!!

This is exactly the kind of tooling gap that slows everything down. I've been running agentic coding setups for months and the number of times I had to manually paste terminal output was absurd. The agent needs to see what you see. Bookmarking this.

sick, looks really good. any plan for adding client logs from browsers as well?

@saiembey Definitely!👀

sick mate!! 🔥

Thanks bro!!

big man fucking cooked

Thank you sir🫡

This is awesome, I was just going to build this.

Thanks!!

This is good. There is a lot of gap for agenticfiicatiin of lot of dev tooling. Good stuff.

Thanks!!

This looks great, gotta give it a go

Thanks!!

That sounds like such a huge productivity booster! I can totally imagine how frustrating it must be to have your agent unaware of local dev activity.

@charlielamb is your terminal config and mac background shared online anywhere?

Not rn, main things im using in this are aerospace and ghostty - background is with backdrop

legend!

Thank you haha!!

works with turbo?

Should do! Try and lmk mate

Love the design 😁

Thank you!!

your exposing unnecessary context rot

Agents can check logs without any tools. Claude Code does it without even asking.

neat!

Thanks!

The agent context gap is brutal. I've wasted hours debugging locally while the AI sits there blind to what's actually happening in my terminal. This fixes the exact friction point that kills productivity.

🙏🙏🚀

so close to build a same thing. I definitely will give this a try.

Cheers! Lmk how u get on!

Cooking 🪵🪵

Thank you bro🤝

Cool! I wanna try it

Lmk how u get on!