grace

Name: grace
Author: Gabriella439

608stars45forks

How to install grace?

gabriella439/grace

$git clone https://github.com/Gabriella439/grace

Installs into the current project.

Summary

Grace (short for Fall-from-Grace) is a domain-specific programming language for prompting models. In particular, Grace is well-suited for building and auto-generating elaborate prompt chains

Files · 1

View on GitHub

README.md

1	# Grace
2
3	[![built with garnix](https://img.shields.io/endpoint.svg?url=https%3A%2F%2Fgarnix.io%2Fapi%2Fbadges%2FGabriella439%2Fgrace)](https://garnix.io/repo/Gabriella439/grace)
4	[![BlueSky \| @trygrace.dev](https://img.shields.io/badge/Bluesky-@trygrace.dev-blue)](https://bsky.app/profile/trygrace.dev)
5
6	Grace (short for [Fall-from-Grace](#name)) is a domain-specific programming
7	language for prompting models. In particular, Grace is well-suited for building
8	and auto-generating elaborate prompt chains
9
10	You can use Grace in your browser without installing anything by visiting
11	[trygrace.dev](https://trygrace.dev/). That website includes an interactive
12	tutorial and is the recommended way to both learn and get started with using
13	Grace.
14
15	## Features
16
17	- Integrated language support for models
18
19	You don't need to install any dependencies or import anything to get started.
20	Everything you need is built directly into the language.
21
22	The language provides a built-in `prompt` function for prompting a model:
23
24	```haskell
25	>>> let key = ./openai.key : Key
26
27	>>> prompt{ key, text: "Generate a list of names" }
28	"
29	Here are 40 varied first names (mixed genders and cultures):
30
31	- Aiden
32	- Sofia
33	- Mateo
34	…
35	- Mabel
36	- Imani
37	- Zane
38
39	Want names filtered by gender, culture, style (modern/vintage), or as full names/surnames?"
40	```
41
42	… and you can structure the output by giving a type annotation:
43
44	```haskell
45	>>> prompt{ key, text: "Generate a list of names" } : List Text
46	[ "Ava Thompson"
47	, "Liam Patel"
48	, "Sophia Martinez"
49	…
50	, "Jackson Rivera"
51	, "Zoe Wilson"
52	, "Aiden Park"
53	]
54	```
55
56	If the type is sufficiently self-explanatory, you can even omit the prompt:
57
58	```haskell
59	>>> prompt{ key } : List { name: Text }
60	[ { "name": "Alice" }
61	, { "name": "Bob" }
62	, { "name": "Charlie" }
63	, { "name": "Diana" }
64	, { "name": "Evan" }
65	]
66	```
67
68	In fact, you can omit the type, too, if the type can be inferred from use:
69
70	```haskell
71	>>> for { name } of prompt{ key } in "Hello, ${name}!"
72	[ "Hello, Alice!"
73	, "Hello, Bob!"
74	, "Hello, Carol!"
75	, "Hello, Dave!"
76	, "Hello, Eve!"
77	]
78	```
79
80	- JSON schemas inferred from use
81
82	That last example works even without a prompt, schema, or type because Grace's
83	type checker reasons backwards from how the output is used to infer the
84	correct JSON schem, like this:
85
86	- the type checker infers that the `name` variable must be `Text`
87
88	… because the `name` variable is interpolated into `"Hello, ${name}!"`
89
90	- the type checker infers that the `prompt` function must generate a `List`
91
92	… because the program loops over the output using a `for … of` loop.
93
94	- the type checker infers each element of the `List` has type `{ name: Text }`
95
96	… because the `for … of` loop destructures each element using `{ name }`
97
98	- therefore the `prompt` function outputs a value of type `List{ name: Text }`
99
100	… which you can read as "a `List` of records, each of which has a `name` field
101	containing `Text`".
102
103	The interpreter then converts that Grace type into the following matching JSON
104	schema to constrain the model's output:
105
106	```json
107	{
108	"type": "array",
109	"items": {
110	"type": "object",
111	"properties": {
112	"name": {
113	"type": "string"
114	}
115	},
116	"required": ["name"],
117	"additionalProperties": false
118	}
119	}
120	```
121
122	Finally, the model infers from that JSON schema alone (without any additional
123	prompt) that it should generate a JSON-encoded list of names.
124
125	- Code generation
126
127	You can prefix the `prompt` keyword with `import` to ask the model to generate
128	a Grace expression of any type. For example:
129
130	```haskell
131	>>> import prompt{ key, text: "increment" }
132	\n -> n + 1
133	```
134
135	You can use an explicit type annotation to guide the generated code:
136
137	```haskell
138	>>> import prompt{ key, text: "increment" } : { input: Natural } -> { output: Natural }
139	\{ input } -> { "output": input + 1 }
140	```
141
142	… and if the type is informa

People who liked this also viewed

6 picks

Type

tunellm chat in text filePromptsJul 202653328
snippetsSupercharge your Snippets in VS Code — Manage your code snippets & AI prompts without quitting your editor.PromptsJul 2026302169
prompt-optimizerAn AI prompt optimizer for writing better prompts and getting better AI results.PromptsJul 2026—32k
chatgpt_system_promptA collection of GPT system prompts and various prompt injection/leaking knowledge.PromptsJul 2026—11k
prompt-masterA Claude skill that writes the accurate prompts for any AI tool. Zero tokens or credits wasted. Full context and memory retentionPromptsJun 2026—11k
python-prompt-toolkitLibrary for building powerful interactive command line applications in PythonPromptsMay 2026—11k

grace

608stars45forks

View on GitHub

How to install grace?

gabriella439/grace

$git clone https://github.com/Gabriella439/grace

Installs into the current project.

Summary

Grace (short for Fall-from-Grace) is a domain-specific programming language for prompting models. In particular, Grace is well-suited for building and auto-generating elaborate prompt chains

Files · 1

View on GitHub

README.md

1	# Grace
2
3	[![built with garnix](https://img.shields.io/endpoint.svg?url=https%3A%2F%2Fgarnix.io%2Fapi%2Fbadges%2FGabriella439%2Fgrace)](https://garnix.io/repo/Gabriella439/grace)
4	[![BlueSky \| @trygrace.dev](https://img.shields.io/badge/Bluesky-@trygrace.dev-blue)](https://bsky.app/profile/trygrace.dev)
5
6	Grace (short for [Fall-from-Grace](#name)) is a domain-specific programming
7	language for prompting models. In particular, Grace is well-suited for building
8	and auto-generating elaborate prompt chains
9
10	You can use Grace in your browser without installing anything by visiting
11	[trygrace.dev](https://trygrace.dev/). That website includes an interactive
12	tutorial and is the recommended way to both learn and get started with using
13	Grace.
14
15	## Features
16
17	- Integrated language support for models
18
19	You don't need to install any dependencies or import anything to get started.
20	Everything you need is built directly into the language.
21
22	The language provides a built-in `prompt` function for prompting a model:
23
24	```haskell
25	>>> let key = ./openai.key : Key
26
27	>>> prompt{ key, text: "Generate a list of names" }
28	"
29	Here are 40 varied first names (mixed genders and cultures):
30
31	- Aiden
32	- Sofia
33	- Mateo
34	…
35	- Mabel
36	- Imani
37	- Zane
38
39	Want names filtered by gender, culture, style (modern/vintage), or as full names/surnames?"
40	```
41
42	… and you can structure the output by giving a type annotation:
43
44	```haskell
45	>>> prompt{ key, text: "Generate a list of names" } : List Text
46	[ "Ava Thompson"
47	, "Liam Patel"
48	, "Sophia Martinez"
49	…
50	, "Jackson Rivera"
51	, "Zoe Wilson"
52	, "Aiden Park"
53	]
54	```
55
56	If the type is sufficiently self-explanatory, you can even omit the prompt:
57
58	```haskell
59	>>> prompt{ key } : List { name: Text }
60	[ { "name": "Alice" }
61	, { "name": "Bob" }
62	, { "name": "Charlie" }
63	, { "name": "Diana" }
64	, { "name": "Evan" }
65	]
66	```
67
68	In fact, you can omit the type, too, if the type can be inferred from use:
69
70	```haskell
71	>>> for { name } of prompt{ key } in "Hello, ${name}!"
72	[ "Hello, Alice!"
73	, "Hello, Bob!"
74	, "Hello, Carol!"
75	, "Hello, Dave!"
76	, "Hello, Eve!"
77	]
78	```
79
80	- JSON schemas inferred from use
81
82	That last example works even without a prompt, schema, or type because Grace's
83	type checker reasons backwards from how the output is used to infer the
84	correct JSON schem, like this:
85
86	- the type checker infers that the `name` variable must be `Text`
87
88	… because the `name` variable is interpolated into `"Hello, ${name}!"`
89
90	- the type checker infers that the `prompt` function must generate a `List`
91
92	… because the program loops over the output using a `for … of` loop.
93
94	- the type checker infers each element of the `List` has type `{ name: Text }`
95
96	… because the `for … of` loop destructures each element using `{ name }`
97
98	- therefore the `prompt` function outputs a value of type `List{ name: Text }`
99
100	… which you can read as "a `List` of records, each of which has a `name` field
101	containing `Text`".
102
103	The interpreter then converts that Grace type into the following matching JSON
104	schema to constrain the model's output:
105
106	```json
107	{
108	"type": "array",
109	"items": {
110	"type": "object",
111	"properties": {
112	"name": {
113	"type": "string"
114	}
115	},
116	"required": ["name"],
117	"additionalProperties": false
118	}
119	}
120	```
121
122	Finally, the model infers from that JSON schema alone (without any additional
123	prompt) that it should generate a JSON-encoded list of names.
124
125	- Code generation
126
127	You can prefix the `prompt` keyword with `import` to ask the model to generate
128	a Grace expression of any type. For example:
129
130	```haskell
131	>>> import prompt{ key, text: "increment" }
132	\n -> n + 1
133	```
134
135	You can use an explicit type annotation to guide the generated code:
136
137	```haskell
138	>>> import prompt{ key, text: "increment" } : { input: Natural } -> { output: Natural }
139	\{ input } -> { "output": input + 1 }
140	```
141
142	… and if the type is informa

People who liked this also viewed

6 picks

Type