Bullet is a coding agent built for speed. We got tired of burning hours waiting on agent runs. The models were fine, but the loops around them were slow. Bullet auto-picks the right model/reasoning level per prompt, parallelizes searches/reads/commands, and uses targeted code search instead of embedding your whole repo. Works with your Claude Code or Codex subscription, API keys, or an on-device model. 95.8% on SWE-bench Verified (top 3), 119s/task. Built by Yale CS grads, ex-AppLovin/Citadel.
Hey everyone, my name is Alex Sima and I am the CTO of Bullet.
Like the post said, we built Bullet because Claude Code and Codex were too slow. Our YC partner said we should solve a problem we face ourselves, so we decided to go big or go home (guess we decided to go small and lightweight but ok).
We now only use Bullet internally...and personally, I can't go back to using Claude Code. If you hate slow coding agents, join the club: https://www.codewithbullet.com.
You can bring your subscriptions over, so it's free speed -- enjoy :)
Would love to see more screenshots of the GUI and how subscriptions can be brought over vs API routing. Settings are also of interest to me to see how customizable this thing is.
I've been using Claude Code CLI for a long time now (since beta days), so I have all of my scaffolding in place. As you know, old habits die hard so it's hard for me to just make the jump and bite the bullet, pun intended. I'm convinced it's unbelievably fast compared to my setup, but my setup does so much of the work for me that it's hard to compare the delta between both...
I'm so excited for this! Recently I've been noticing that the bottleneck with coding agents isn't quality anymore, it's just speed - so excited to have bullet accelerating my dev velocity!
Stumbled across this post as my claude code was taking 30 years. How timely
I would love to download and start using this today, but can't justify burning tokens when claude code is $200 a month and I spend $3-4k worth of tokens :(
Congrats! This is so cool :) Quick question — besides the Minecraft clone example in the video, what are some other real-world coding tasks where you think Bullet’s speed improvements would make the biggest difference?
Congrats on the launch@alex_sima26@adi_kulkarni! Does Bullet also help stretch usage limits, or is it just speed? Any benchmarks or stats on token savings?
Congrats on the launch! Waiting on coding agents is easily one of the most annoying parts right now, so cutting down iteration time is a big win.
congrats on the launch guys! looks very impressive, excited to see where this goes!!!
Hi everyone, my name is @adi_kulkarni and I’m the CEO of Bullet.
Existing coding agents were too slow for our development workflows, especially when iterating on a feature synchronously. Even when they finished, there would be a mistake so countless hours were wasted in this loop. After a certain amount of context accumulates in a session, both the agent and I are confused about the actual state of the feature!
Thus, we built Bullet to make coding agents faster, more focused, and easier to iterate with. It automatically chooses the right model and reasoning level, runs independent work in parallel, and uses targeted code search instead of dragging unnecessary context through every step.
We built Bullet for ourselves first, and now we’re excited to share it with everyone. Would love to hear what you all think!
About Bullet on Product Hunt
“30-60% faster than Claude Code and Codex”
Bullet launched on Product Hunt on August 11th, 2026 and earned 227 upvotes and 38 comments, placing #5 on the daily leaderboard. Bullet is a coding agent built for speed. We got tired of burning hours waiting on agent runs. The models were fine, but the loops around them were slow. Bullet auto-picks the right model/reasoning level per prompt, parallelizes searches/reads/commands, and uses targeted code search instead of embedding your whole repo. Works with your Claude Code or Codex subscription, API keys, or an on-device model. 95.8% on SWE-bench Verified (top 3), 119s/task. Built by Yale CS grads, ex-AppLovin/Citadel.
Bullet was featured in Productivity (659.5k followers), Developer Tools (518.4k followers) and Vibe coding (646 followers) on Product Hunt. Together, these topics include over 239.4k products, making this a competitive space to launch in.
Who hunted Bullet?
Bullet was hunted by Garry Tan. A “hunter” on Product Hunt is the community member who submits a product to the platform — uploading the images, the link, and tagging the makers behind it. Hunters typically write the first comment explaining why a product is worth attention, and their followers are notified the moment they post. Around 79% of featured launches on Product Hunt are self-hunted by their makers, but a well-known hunter still acts as a signal of quality to the rest of the community. See the full all-time top hunters leaderboard to discover who is shaping the Product Hunt ecosystem.
Want to see how Bullet stacked up against nearby launches in real time? Check out the live launch dashboard for upvote speed charts, proximity comparisons, and more analytics.
Garry Tan
Hey everyone, my name is Alex Sima and I am the CTO of Bullet.
Like the post said, we built Bullet because Claude Code and Codex were too slow. Our YC partner said we should solve a problem we face ourselves, so we decided to go big or go home (guess we decided to go small and lightweight but ok).
We now only use Bullet internally...and personally, I can't go back to using Claude Code. If you hate slow coding agents, join the club: https://www.codewithbullet.com.
You can bring your subscriptions over, so it's free speed -- enjoy :)