TechWorks

Building Fluxen 🚀
AI • SaaS • Entrepreneurship
Building in public. Sharing what I learn along the way.

[TechWorks @techworks9999]

@anupamrjp https://t.co/RbkAz8zu0g

[TechWorks @techworks9999]

@mytwillot Good one! Will be definitely useful

[TechWorks @techworks9999]

What are you building this weekend!

#saas #founder

[TechWorks @techworks9999]

Which is better copilot or claude or codex?

#llm

[TechWorks @techworks9999]

@Mudhaxk Done follow back!

[TechWorks @techworks9999]

Hey founders! 🚀

Looking to connect with people building in:

📦 SaaS
🤖 AI Tools
🏧 Automation
🧑‍💻 Dev
📈 Marketing
💡 Solopreneurs

Let's #connect and grow together this weekend!

[TechWorks @techworks9999]

@syed_shahmir5 Done great connecting with you

[TechWorks @techworks9999]

@VishuLIV Done follow back!

[TechWorks @techworks9999]

@ShaviPathania Done great connecting with you!

[TechWorks @techworks9999]

@s7anit Great to know and great connecting with you!

[TechWorks @techworks9999]

@devcoderaditya Followed! Follow back!

[TechWorks @techworks9999]

@asaridarr Done great connecting with you!

[TechWorks @techworks9999]

Any one use LLM APIs for their saas or any other application?

How you guys are managing costs?

[TechWorks @techworks9999]

@decodewithrudra Lets connect! Follow back

[TechWorks @techworks9999]

@Grubbyzbc Followed back

[TechWorks @techworks9999]

@itskamrankhan Lets connect followed you! Follow back

[TechWorks @techworks9999]

@openingai_com Done! please follow back

[TechWorks @techworks9999]

@trn3n Followed nice connecting with you!!

[TechWorks @techworks9999]

wtf are you guys building right now drop it down so it get some eyes.

[TechWorks @techworks9999]

@MaxshotAI Good breakdown. The gap we kept hitting: gateways show you the data but leave the decisions to you. That's why we built Fluxen (open source) to flag waste, simulate fixes against real traffic, and check whether they actually worked.
https://t.co/RbkAz8zu0g

[TechWorks @techworks9999]

Soft alert at 80%, hard cap at 100%, per team or agent. Hitting the cap pauses non-critical work, and a lead approves temporary bumps. What makes it work is reviewing every bump. Otherwise the limit just ratchets up.

We're building this into Fluxen, an open-source AI gateway: each app gets its own budget and rate limits, and it flags where spend is being wasted so you can lower caps instead of raising them. https://t.co/Sk4WZkCnrv