BAML is an expressive language for structured text generation
$ pip install baml-py
Backed by
Trusted by developers at
Forget JSON.parse(). BAML's parser was built for LLMs
LLM yapping immunity
JSON error correction
Schema coercion
BAML automatically finds your schema amongst any other additional generated text. Use techniques like chain-of-thought with ease no matter how complex your schema is.
LLM output
Parsed Response
Function-calling for every model, in your favorite language
All models get better with BAML
BAML achieves state-of-the-art results in function-calling with GPT 3.5 over nearly all other models and frameworks.
.4%
AST Score
+
.9%
Gain for GPT4o
Berkeley Function-Calling Benchmark (AST Score)
BAML was designed for easier, more reliable prompting
Syntax that transforms prompt engineering to coding.
file_name.baml
Developers love building with BAML
BAML is amazing. I've used it in Python and Typescript. It's a game changer.
Just set up baml for my project, 10/10 experience and much faster than langchain.
Finic.ai
It's amazing!! Was able to cut down my tokens and time-to-first-token significantly without compromising results.
Cerebral Valley
BAML is definitely a must have if you want any structured data from LLM; no more BS/long paragraphs describing what the output should be like, it just works!!!
Coldreach.ai
BAML is amazing. I've used it in Python and Typescript. It's a game changer.
Just set up baml for my project, 10/10 experience and much faster than langchain.
Finic.ai
It's amazing!! Was able to cut down my tokens and time-to-first-token significantly without compromising results.
Cerebral Valley
BAML is definitely a must have if you want any structured data from LLM; no more BS/long paragraphs describing what the output should be like, it just works!!!
Coldreach.ai
BAML is amazing. I've used it in Python and Typescript. It's a game changer.
Just set up baml for my project, 10/10 experience and much faster than langchain.
Finic.ai
It's amazing!! Was able to cut down my tokens and time-to-first-token significantly without compromising results.
Cerebral Valley
BAML is definitely a must have if you want any structured data from LLM; no more BS/long paragraphs describing what the output should be like, it just works!!!
Coldreach.ai
BAML is amazing. I've used it in Python and Typescript. It's a game changer.
Just set up baml for my project, 10/10 experience and much faster than langchain.
Finic.ai
It's amazing!! Was able to cut down my tokens and time-to-first-token significantly without compromising results.
Cerebral Valley
BAML is definitely a must have if you want any structured data from LLM; no more BS/long paragraphs describing what the output should be like, it just works!!!
Coldreach.ai
The test case and playground is quite literally the BEST feature. It has improved the iteration speed and quality by an order of magnitude.
Docucare AI
I really really like what Baml offers [...] I think it's a step-wise improvement over Marvin. Having complete control over the prompt WITH strong type guarantees is fantastic.
I also think the dedicated testing playground is awesome.
I also think the dedicated testing playground is awesome.
Zenfetch
Code is hella clean now. Look at [the] folder structure, and each folder for a respective pipeline. Each file just a prompt. Clean, elegant, beautiful.
Magnaplay
Just got the categorizer to work first try.
Felt like landing a kickflip
Felt like landing a kickflip
Squack
The test case and playground is quite literally the BEST feature. It has improved the iteration speed and quality by an order of magnitude.
Docucare AI
I really really like what Baml offers [...] I think it's a step-wise improvement over Marvin. Having complete control over the prompt WITH strong type guarantees is fantastic.
I also think the dedicated testing playground is awesome.
I also think the dedicated testing playground is awesome.
Zenfetch
Code is hella clean now. Look at [the] folder structure, and each folder for a respective pipeline. Each file just a prompt. Clean, elegant, beautiful.
Magnaplay
Just got the categorizer to work first try.
Felt like landing a kickflip
Felt like landing a kickflip
Squack
The test case and playground is quite literally the BEST feature. It has improved the iteration speed and quality by an order of magnitude.
Docucare AI
I really really like what Baml offers [...] I think it's a step-wise improvement over Marvin. Having complete control over the prompt WITH strong type guarantees is fantastic.
I also think the dedicated testing playground is awesome.
I also think the dedicated testing playground is awesome.
Zenfetch
Code is hella clean now. Look at [the] folder structure, and each folder for a respective pipeline. Each file just a prompt. Clean, elegant, beautiful.
Magnaplay
Just got the categorizer to work first try.
Felt like landing a kickflip
Felt like landing a kickflip
Squack
The test case and playground is quite literally the BEST feature. It has improved the iteration speed and quality by an order of magnitude.
Docucare AI
I really really like what Baml offers [...] I think it's a step-wise improvement over Marvin. Having complete control over the prompt WITH strong type guarantees is fantastic.
I also think the dedicated testing playground is awesome.
I also think the dedicated testing playground is awesome.
Zenfetch
Code is hella clean now. Look at [the] folder structure, and each folder for a respective pipeline. Each file just a prompt. Clean, elegant, beautiful.
Magnaplay
Just got the categorizer to work first try.
Felt like landing a kickflip
Felt like landing a kickflip
Squack
Get high quality results
Talk to one of our AI Experts
We have seen hundreds of prompts and AI pipelines. Talk to a founder to learn more about our roadmap, our enterprise support and observability platform, or join our community to see what others are up to.