This product was not featured by Product Hunt yet. It will not be visible on their landing page and won't be ranked (cannot win product of the day regardless of upvotes).
Product upvotes vs the next 3
Waiting for data. Loading
Product comments vs the next 3
Waiting for data. Loading
Product upvote speed vs the next 3
Waiting for data. Loading
Product upvotes and comments
Waiting for data. Loading
Product vs the next 3
Loading
Haku
Mac + browser automation for AI agents — 85% fewer tokens
Haku gives your AI agents seamless control of your Mac — native apps, Chrome, Safari — through one MCP server. An on-device model filters each page and app down to the few elements your agent actually needs(In milliseconds). As a result, a chrome action drops from the usual 50K tokens to under 7K. Works with Claude Code, Codex, OpenClaw, Cursor, or any MCP client. Fully local. It also ships with agent-driven demo recordings that lets your agent create beautiful demos and onboarding videos
Haku is 3 products rolled into one, and it is the culmination of 3 of my major obsessions.
1. Mac automation. I have been pursuing this since the time of early GPT-4. At some point I was combining AppleScript with the number of tab keys required to reach a click target :D Super hacky. But with better models, thankfully, I don't have to do that anymore. With Haku, Mac automation finally feels seamless - on par with Claude's computer use, but at a fraction of the token spend.
2. Frontend testing. I absolutely hate clicking around my own apps to test them, and will avoid it at any cost. Every time I tested with Chrome DevTools or Claude in Chrome, I'd watch agents spend 40k, 70k tokens on just a few flows. I wanted it cheaper, and that is how the filter was born. Two extremely fast filters _ one of them a super tiny classifier running on the Neural Engine _ sit between what CDP sends and what Claude sees, cutting context down to roughly a fifth of what agents would normally see. It started off with this, and then extended to full browser automation. Amazon, LinkedIn, X, Swiggy (food delivery app in India) _ with Haku, my agents can fully navigate and do tasks on any of them.
3. Token consumption. This became even more important with the OpenClaw release. It is amazing, but it was eating away way too many tokens. The same filter I built for the web saved me here too _ one idea, three surfaces.
I hope you find Haku as useful and as transformative as I do. If any of these three obsessions are yours too, I'd love to hear what you automate with it.
About Haku on Product Hunt
“Mac + browser automation for AI agents — 85% fewer tokens”
Haku was submitted on Product Hunt and earned 6 upvotes and 1 comments, placing #60 on the daily leaderboard. Haku gives your AI agents seamless control of your Mac — native apps, Chrome, Safari — through one MCP server. An on-device model filters each page and app down to the few elements your agent actually needs(In milliseconds). As a result, a chrome action drops from the usual 50K tokens to under 7K. Works with Claude Code, Codex, OpenClaw, Cursor, or any MCP client. Fully local. It also ships with agent-driven demo recordings that lets your agent create beautiful demos and onboarding videos
On the analytics side, Haku competes within Mac and Artificial Intelligence — topics that collectively have 570.8k followers on Product Hunt. The dashboard above tracks how Haku performed against the three products that launched closest to it on the same day.
Who hunted Haku?
Haku was hunted by Shahir M A. 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.
For a complete overview of Haku including community comment highlights and product details, visit the product overview.
Shahir M A
Haku is 3 products rolled into one, and it is the culmination of 3 of my major obsessions.
1. Mac automation. I have been pursuing this since the time of early GPT-4. At some point I was combining AppleScript with the number of tab keys required to reach a click target :D Super hacky. But with better models, thankfully, I don't have to do that anymore. With Haku, Mac automation finally feels seamless - on par with Claude's computer use, but at a fraction of the token spend.
2. Frontend testing. I absolutely hate clicking around my own apps to test them, and will avoid it at any cost. Every time I tested with Chrome DevTools or Claude in Chrome, I'd watch agents spend 40k, 70k tokens on just a few flows. I wanted it cheaper, and that is how the filter was born. Two extremely fast filters _ one of them a super tiny classifier running on the Neural Engine _ sit between what CDP sends and what Claude sees, cutting context down to roughly a fifth of what agents would normally see. It started off with this, and then extended to full browser automation. Amazon, LinkedIn, X, Swiggy (food delivery app in India) _ with Haku, my agents can fully navigate and do tasks on any of them.
3. Token consumption. This became even more important with the OpenClaw release. It is amazing, but it was eating away way too many tokens. The same filter I built for the web saved me here too _ one idea, three surfaces.
I hope you find Haku as useful and as transformative as I do. If any of these three obsessions are yours too, I'd love to hear what you automate with it.