Forge

The opinionated stack for AI copilots

Built @fractaltechnyc

[Forge @ml_forge]

Shoutout to the real ones

[los (appstar) @downloadlos]

another day using @ml_forge. its beautiful

[Forge @ml_forge]

@downloadlos 🫡

[Forge @ml_forge]

@downloadlos @aphysicist RAG that bad boy

[Forge @ml_forge]

@downloadlos 🙏

[Fractal Tech @fractaltechnyc]

when you are procedurally generating structured data with @ml_forge using canned prompt templates on autopilot

[Forge @ml_forge]

First line of code to production in <24hrs

props to @mehulbehera0 for the build!

[Jake Zegil 🗽 @JakeZegil]

Launching BestieChats today - powered by @ml_forge

Chat with your favorite characters!

[Jake Zegil 🗽 @JakeZegil]

type safety is 10x more powerful with AI (less overhead) and AI is 10x more powerful with types (more guardrails)

[Jake Zegil 🗽 @JakeZegil]

Forge one-line RAG is now in beta 🚀

If you, too, are sick of managing context and wrestling with llamaindex for agent context, just upload docs to forge and prompt. Support for Groq, OpenAI, and Anthropic.

Sign up in comments : )

[Forge @ml_forge]

@nisreenmandvi @ethansteininger @JakeZegil We're working on it : )

[Nisreen @nisreenmandvi]

Built with @ml_forge in 5 minutes

[Forge @ml_forge]

ahem, dead simple

[Jake Zegil 🗽 @JakeZegil]

@ml_forge now supports Groq

And it's pretty freaking fast

[Jake Zegil 🗽 @JakeZegil]

Cursor is an insanely good example of why RAG isn't going anywhere

The suggestion context is extremely well curated and FAST, and the results speak for themselves

[Jake Zegil 🗽 @JakeZegil]

On Sunday, one of my students sent me of my favorite @ml_forge use case so far - the easiest i18n library you've ever used

[Jake Zegil 🗽 @JakeZegil]

Onboarding a user to forge today:

"Wait, it's that easy??"

[Forge @ml_forge]

Forge + Claude is about to get very fast

[Alex Albert @alexalbert__]

We just rolled out prompt caching in the Anthropic API.

It cuts API input costs by up to 90% and reduces latency by up to 80%.

Here's how it works:

[Forge @ml_forge]

@alexalbert__ Going to roll this out for Anthropic by EOD today - input tokens for guaranteed structured data are highly repetitive

This will make Anthropic via Forge significantly more cost effective than OpenAI

[Forge @ml_forge]

Big day for us - we've partnered with @getlsd to bring your AI products online 🌎

Oh yeah, we also went multi-provider, launching @AnthropicAI models!

[Jake Zegil 🗽 @JakeZegil]

I built out this AI-powered skin of hacker news in about 20m with @ml_forge and @getlsd

It uses live data from https://t.co/5GKP7icArV

Killer combo for building online AI products that use realtime web data

[Jake Zegil 🗽 @JakeZegil]

@virattt @ml_forge guarantees it across all models (and faster), including 4o-mini with images. Also way better dev exp

[Mehul Behera @mehulbehera0]

@AlphaSignalAI its cool but I had a way better dev experience w @ml_forge