๐ FFrames goes beta ๐
Already every rust (and not only rust) developer can join our public beta testing and try out video production with @fframes_rust
I am thrilled and super excited on how awesome and powerful the project became already and there is much more to come.
My friends at XCrawl paid me some good money to tell you about their API cause it's great. So you better be reading this carefully:
A lot of folks today and likely you are building or going to build some agentic workflows that need to be smart but they came out looking dumb as they lack *REAL* information about the world.
You will try to make your own %WebSearch% tool with your clanker and will make a rookie mistake. Just hear me out: clankers love writing code, but what they actually NEED is a clean trusted source of information to populate the context, not a playwright script that fails 3/10 times.
What you really want for a smart agentic workflow is just a reliable service that will give your AI a context of the web in the format it understands.
And this is XCrawl.
like for real imagine you want to constantly track https://t.co/PI6MX4KZ4t for data. With XCrawl you literally do `curl xcrawl/justfuckinguseneovim.com` and it returns the full website in markdown, json, or even sitemap format -> you feed this to your clanker -> it gets shit done.
If you are going to build it yourself via playwright, you are going to have thousands loc, will require a runtime, deployment, will need to maintain a browser and will still fail cause they my website has a naive bot protection based on IP masking. With XCrawl you just get the *CLEAN* markdown with the exact website content each time you request the data.
Oh and they asked me to mention that they have an industry-leading anti-blocking with over 90% success rate. They promise full ethics and compliance - no malicious activity, access only to public data and accuracy.
AI agents searching the web aren't dumb, they just look dumb cause reading the web is not trivial. So if you are building some AI workflows based on the web context check out XCrawl.
Here is a link with some free credit so you can test it out yourself https://t.co/Z75wAS6NZq
my friends at @recallai are hosting the meetup about fff in SF
come to meet out, get some food, I will make a small talk and we can just hang out
https://t.co/47esVmWeTB
It's pretty dishonest that software like ripgrep considered "free" just because folks never see a memory and CPU usage impact when it's executing as a child process
but it's not free
On every query in chromium repo it allocates 58 megabytes and what is worse it needs to always return it back to the system and allocate again.
every single time you search. again and again
My dear software engineers, I am excited to present you my latest achievement in the code search area that I've been trying to tackle for the last months:
ACTUALLY WORKING real-time approximate typo-resistant code search. What does it mean?
you: can search any code with any typos
you agent: for every search of UserController with 0 results will automatically suggest UserAuthController without additional cost
It's already live https://t.co/5X6nOmdf5r you can try it right now as MCP for file search
My dear software developers (and anyone whoโs interested in the future of code search):
I have crawled through depths of hell to bring you, one of the more important foundational piece of programming: fast, the most accurate, index free, and correct code search
Here is a real time code search on leaked claude code sources, linux kernel 100k files, and chromium repo 500k files
๐จ BREAKING: fframes studio laying off 167 employees and shutting down, following FFmpeg announcement of moving to Rust.
"There is no value proposition to our product anymore", says fframes CEO
He then added:
"FFF is the fastest file search in the world"
Let's just all agree that
curl -L <url>.sh | bash
is just the best way to install any software. It's not even worth to discuss. No one wants to use the shitty package manager especially linux users
just already make a script and make a dns redirect to github. That's all.
Introducing fff-ai. It's a file search tooling optimized specifically for your AI
1) significantly faster than fzf and ripgrep
2) has fuzzy code search fallbacks
3) better sort and suggestions of access frecency, git status, file size, etc
Avg -10% wall time and -17% tokens
I find it so dumb that I made a framework that can render and encode 30 seconds full-hd video with every unique frame *ON CPU* in 1s (7-8s real time)
and still didn't publish it
This is why fframes provides both renderers backend for SVG: SIMD optimized cpu and gpu.
CPU rendering is really cool when you donโt ship to the end user machine
GPU is just cool