this is pure f*cking treasure 15 GitHub projects with 1.21M combined stars that can form a real agent stack specs. memory. web data. documents. context. sandboxes. monitoring. video 01 hermes-agent ▸ 02 OpenSpec ▸ 03 caveman ▸ 04 Scrapling ▸ 05 Docling ▸ 06 PageIndex ▸ 07 mem0 ▸... 08 headroom ▸ 09 Daytona ▸ 10 TrendRadar ▸ 11 Fabric ▸ 12 spec-kit ▸ 13 hyperframes ▸ 14 OpenMontage ▸ 15 AI Engineering Hub ▸ the loop: define the job → collect the evidence → parse the docs → save the memory → compress the context → run the code safely → watch what changes → ship the output the entire stack is open source save this to build your own business with the help of an AI employee ⭣show more

cool list but holy visual larp slop 😭

how many of them do you use on a daily basis? and which one do you think is the best?

docling is the sleeper here.

@beamnxw the memory part is such a game changer. while building my stuff, I realized it’s crucial for real context handling.

gh is such an amazing place, honestly don’t know what isn’t on there

i already have hermes the rest of the list i'll leave

It's worth getting this kind of good stuff every night and paying for it 😄

Wow, these are some really cool repo

the loop in 8 verbs: define, collect, parse, save, compress, run, watch, ship. harness without a vendor

the fact that mem0 and Fabric show up here is huge. context compression is what actually determines if agents stay useful at scale

15 gems on the way

Bookmarking this, thanks for compiling

I'm amazed how you constantly research them?

saved this list

mem0 always boosted my context recall speed

There goes the weekend.

I save it absolutely banger

Agreed

This is a useful way to think about agents as a stack instead of a single magic model. The part I would add is failure visibility. Memory, sandboxes, and monitoring only help if the system makes its uncertainty easy to inspect. Which layer do you think is still weakest?

open-source components still need orchestration to work as one stack

The loop is the stack. The repos are just the parts

saved for tomorrow :)

fair agree, curating open source primitives into a complete agent stack saves months of stitching things together from scratch, letting developers focus on shipping actual features instead of reinventing plumbing

Love the stack. Add a budget per run early - retry loops are where “cheap” agents get expensive 😅

mapping the workflow gives each project a defined role in composition.

i have been looking for something like this for a long time thank you

docling and mem0 alone make this list worth it

@threadreaderapp unroll

Specs and skills beat vibes. Put the contract in the repo before the agent touches code. Curious what you version alongside the skill so it does not rot.

agents very nice cooking

top repos