https://github.com/hellerve/llml
A silly programming language that lets an LLM infer
https://github.com/hellerve/llml
Last synced: 6 months ago
JSON representation
A silly programming language that lets an LLM infer
- Host: GitHub
- URL: https://github.com/hellerve/llml
- Owner: hellerve
- Created: 2025-01-15T10:27:55.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2025-05-28T10:04:34.000Z (10 months ago)
- Last Synced: 2025-08-13T11:18:50.853Z (7 months ago)
- Language: Smalltalk
- Size: 19.5 KB
- Stars: 1
- Watchers: 1
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# llml
A silly language that lets LLMs do the heavy lifting.
You can write code yourself:
```st
LlmlEvaluator new evaluate:
(LlmlParser parse:
'(def x (fn (n) (* n 2)))
(+ (x 10) 3)') "=> 23"
```
Or instead let an LLM do the heavy lifting:
```st
LlmlEvaluator new evaluate:
(LlmlParser parse:
'(def factorial (infer))
(factorial 6)') "=> 720"
```
The inference uses your default LLM connection as configured inside Glamorous Toolkit. You can also select one by hand:
```
LlmlEvaluator new
connection: (GtLlmConnection new
provider: GtOllamaProvider;
model: 'llama3.2');
evaluate: (LlmlParser
parse: '(def factorial (infer))
(factorial 4)')
```
## Installation
```st
Metacello new
repository: 'github://hellerve/llml:main/src';
baseline: 'LlmL';
load
```
Have fun!