rickub
bots-garden/mini-mepublic Fork
⑂ d86eca0
Branches
feature/agents-file feature/ask-agent-hooks feature/ask-cli feature/declared-tools feature/display-called-hooks feature/fetch-build-in feature/genkit-v1.13.1 feature/github-mcp-in-sbx feature/go-1.27.3 feature/hooks feature/improve-init feature/interactive-prompt feature/last-question feature/llmman-provider feature/log-errors feature/lsp-builtin feature/mcp-gateway-session feature/mcp-support feature/memory-search feature/mini-mes feature/mm-init feature/open-terminal feature/prompt-flag feature/response-request feature/resume-acp feature/sbx-template-kit feature/search-in-doc feature/selectors feature/skill-discovery feature/structured-output feature/system-one feature/talk-to-another-mm feature/tasks-builtin feature/try-again feature/unload-model feature/wasm-custom-tools feature/yolo feature/zed-skill-selector fix/generation-timeout fix/plan-mode fix/too-long-completion fix/tool-order fix/word-wrap improve-mem-usage main thinking
Tags
v0.0.0 v0.0.1 v0.0.2 v0.0.3 v0.0.4 v0.1.0 v0.1.1 v0.1.2 v0.1.3 v0.1.4 v0.1.5 v0.1.6 v0.1.7 v0.1.8 v0.1.9 v0.2.0 v0.2.1 v0.2.2 v0.2.3 v0.2.4 v0.2.5 v0.2.6 v0.3.0 v0.4.0 v0.5.0 v0.6.0 v0.7.0 v0.8.0 v0.9.0 v1.0.0 View all branches → View all tags →
Commits
⬇ Clone ▾
git clone https://git.rickub.com/bots-garden/mini-me.git
git clone ssh://git@rickub.com/bots-garden/mini-me.git

Host key fingerprint (ed25519): SHA256:iycHnxEyq0Q7uyVpB7JlznP0G7JrTPXLYRcAU5CSLhc — verify it before your first connect.

Download ZIP

Add the structured-output-pizza demo

One JSON agent declared with Go types (Gemma 4 E4B on llmman), listed in
the README's demo table.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
k33g committed 2026-10-16T09:10:07+02:00 Browse files
d86eca0 parent: 335ef35
modified README.md +1 -0
@@ -129,6 +129,7 @@ The `demos/` directory holds ready-made set-ups (most with a README explaining h
129129 | `demos/llmman/` | mm on llmman, and an agent that asks other agents |
130130 | `demos/ollama/` | mm on Ollama |
131131 | `demos/structured-output/` | agents that answer in JSON, and a coordinator using them |
132+| `demos/structured-output-pizza/` | one JSON agent declared with Go types: a pizza's name → its ingredients (quantity, unit, category, allergens), whether it is vegetarian, its baking time (Gemma 4 E4B on llmman) |
132133 | `demos/agents-md-file/` | an agent following a project's `AGENTS.md` |
133134 | `demos/mini-mes/` | small teams of agents (`riker` coordinating a specifier and a coder, `gopher` a Go specialist) |
134135 | `demos/little-coder/` | a Go + htmx web-app generator for a small local model (Mellum2 12B on llmman, 32 GB RAM): a tested starter, skills, `AGENTS.md`, measured results |
@@ -129,6 +129,7 @@ The `demos/` directory holds ready-made set-ups (most with a README explaining h
129 | `demos/llmman/` | mm on llmman, and an agent that asks other agents |129 | `demos/llmman/` | mm on llmman, and an agent that asks other agents |
130 | `demos/ollama/` | mm on Ollama |130 | `demos/ollama/` | mm on Ollama |
131 | `demos/structured-output/` | agents that answer in JSON, and a coordinator using them |131 | `demos/structured-output/` | agents that answer in JSON, and a coordinator using them |
132+| `demos/structured-output-pizza/` | one JSON agent declared with Go types: a pizza's name → its ingredients (quantity, unit, category, allergens), whether it is vegetarian, its baking time (Gemma 4 E4B on llmman) |
132 | `demos/agents-md-file/` | an agent following a project's `AGENTS.md` |133 | `demos/agents-md-file/` | an agent following a project's `AGENTS.md` |
133 | `demos/mini-mes/` | small teams of agents (`riker` coordinating a specifier and a coder, `gopher` a Go specialist) |134 | `demos/mini-mes/` | small teams of agents (`riker` coordinating a specifier and a coder, `gopher` a Go specialist) |
134 | `demos/little-coder/` | a Go + htmx web-app generator for a small local model (Mellum2 12B on llmman, 32 GB RAM): a tested starter, skills, `AGENTS.md`, measured results |135 | `demos/little-coder/` | a Go + htmx web-app generator for a small local model (Mellum2 12B on llmman, 32 GB RAM): a tested starter, skills, `AGENTS.md`, measured results |
added demos/structured-output-pizza/README.en.md +63 -0
new file mode 100644
@@ -0,0 +1,63 @@
1+# JSON answers from Go types — a pizza's ingredients
2+
3+*[Version française](README.fr.md)*
4+
5+One agent that answers **only in JSON**: name a pizza, get its ingredients for one 30 cm pizza. The shape of the answer is declared as Go types under `structuredOutput` in `agent.yaml`.
6+
7+Model: **Gemma 4 E4B** (`hf.co/unsloth/gemma-4-E4B-it-GGUF:Q4_K_M`) on **llmman**.
8+
9+## Run
10+
11+```bash
12+llmman pull hf.co/unsloth/gemma-4-E4B-it-GGUF:Q4_K_M
13+cd demos/structured-output-pizza
14+mm -p "A four-cheese pizza" | jq .
15+```
16+
17+In prompt mode (`-p`), stdout holds the JSON alone, so it pipes into `jq`.
18+
19+## The declaration
20+
21+```go
22+// One ingredient of the pizza.
23+type Ingredient struct {
24+ Name string `json:"name"` // the ingredient's name, in the user's language
25+ Quantity float64 `json:"quantity"` // amount for one pizza
26+ Unit string `json:"unit"` // "g", "ml", "piece", "pinch" or "to taste"
27+ Category string `json:"category"` // "base", "sauce", "cheese", "meat", "vegetable", "herb" or "other"
28+ Allergens []string `json:"allergens"` // e.g. "gluten", "lactose"; empty when none
29+ Optional bool `json:"optional"`
30+}
31+
32+type Pizza struct {
33+ Name string `json:"name"`
34+ Vegetarian bool `json:"vegetarian"`
35+ Ingredients []Ingredient `json:"ingredients"`
36+ BakingTime *int `json:"bakingTime"` // minutes, null when unknown
37+}
38+
39+var Recipe Pizza
40+```
41+
42+- The `var` names the root of the answer: here one `Pizza` object. `var Recipes = []Pizza{}` would ask for an array.
43+- The `//` comments become the fields' `description` in the JSON Schema; they are where the allowed values go, since `const` (and so enums) is refused.
44+- Every key is required; a pointer (`*int`) also allows `null`.
45+- An agent with `structuredOutput` declares no tool: `bashTool`, `editTools`, `planTool`, `fetchContentTool`, `askUserTool` are `false` and `skillsDir` is `no-skills`.
46+
47+## Seen on 2026-10-16
48+
49+`mm -p "A four-cheese pizza"` (about 40 s):
50+
51+```
52+Four-Cheese Pizza — vegetarian: true — bakingTime: 12
53+Pizza Dough 300 g (base), Tomato Sauce 50 g (sauce), Mozzarella 100 g, Gorgonzola 50 g,
54+Provolone 50 g, Parmesan 20 g (cheese), Olive Oil 15 ml (other)
55+```
56+
57+`mm -p "Une pizza margherita"`: five ingredients, names in Italian (`Mozzarella di bufala`, `Basilico fresco`…) — the model chose the dish's own language over the question's.
58+
59+## Further reading
60+
61+- Every Go construct and its JSON Schema: [configuration reference — `structuredOutput`](../../docs/en/reference/configuration.md#structuredoutput)
62+- Step by step: [Get answers as JSON](../../docs/en/how-to/get-json-answers.md)
63+- Several JSON agents and a coordinator: `demos/structured-output/`
new file mode 100644
@@ -0,0 +1,63 @@
1+# JSON answers from Go types — a pizza's ingredients
2+
3+*[Version française](README.fr.md)*
4+
5+One agent that answers **only in JSON**: name a pizza, get its ingredients for one 30 cm pizza. The shape of the answer is declared as Go types under `structuredOutput` in `agent.yaml`.
6+
7+Model: **Gemma 4 E4B** (`hf.co/unsloth/gemma-4-E4B-it-GGUF:Q4_K_M`) on **llmman**.
8+
9+## Run
10+
11+```bash
12+llmman pull hf.co/unsloth/gemma-4-E4B-it-GGUF:Q4_K_M
13+cd demos/structured-output-pizza
14+mm -p "A four-cheese pizza" | jq .
15+```
16+
17+In prompt mode (`-p`), stdout holds the JSON alone, so it pipes into `jq`.
18+
19+## The declaration
20+
21+```go
22+// One ingredient of the pizza.
23+type Ingredient struct {
24+ Name string `json:"name"` // the ingredient's name, in the user's language
25+ Quantity float64 `json:"quantity"` // amount for one pizza
26+ Unit string `json:"unit"` // "g", "ml", "piece", "pinch" or "to taste"
27+ Category string `json:"category"` // "base", "sauce", "cheese", "meat", "vegetable", "herb" or "other"
28+ Allergens []string `json:"allergens"` // e.g. "gluten", "lactose"; empty when none
29+ Optional bool `json:"optional"`
30+}
31+
32+type Pizza struct {
33+ Name string `json:"name"`
34+ Vegetarian bool `json:"vegetarian"`
35+ Ingredients []Ingredient `json:"ingredients"`
36+ BakingTime *int `json:"bakingTime"` // minutes, null when unknown
37+}
38+
39+var Recipe Pizza
40+```
41+
42+- The `var` names the root of the answer: here one `Pizza` object. `var Recipes = []Pizza{}` would ask for an array.
43+- The `//` comments become the fields' `description` in the JSON Schema; they are where the allowed values go, since `const` (and so enums) is refused.
44+- Every key is required; a pointer (`*int`) also allows `null`.
45+- An agent with `structuredOutput` declares no tool: `bashTool`, `editTools`, `planTool`, `fetchContentTool`, `askUserTool` are `false` and `skillsDir` is `no-skills`.
46+
47+## Seen on 2026-10-16
48+
49+`mm -p "A four-cheese pizza"` (about 40 s):
50+
51+```
52+Four-Cheese Pizza — vegetarian: true — bakingTime: 12
53+Pizza Dough 300 g (base), Tomato Sauce 50 g (sauce), Mozzarella 100 g, Gorgonzola 50 g,
54+Provolone 50 g, Parmesan 20 g (cheese), Olive Oil 15 ml (other)
55+```
56+
57+`mm -p "Une pizza margherita"`: five ingredients, names in Italian (`Mozzarella di bufala`, `Basilico fresco`…) — the model chose the dish's own language over the question's.
58+
59+## Further reading
60+
61+- Every Go construct and its JSON Schema: [configuration reference — `structuredOutput`](../../docs/en/reference/configuration.md#structuredoutput)
62+- Step by step: [Get answers as JSON](../../docs/en/how-to/get-json-answers.md)
63+- Several JSON agents and a coordinator: `demos/structured-output/`
added demos/structured-output-pizza/README.fr.md +57 -0
new file mode 100644
@@ -0,0 +1,57 @@
1+# Des réponses JSON à partir de types Go — les ingrédients d'une pizza
2+
3+*[English version](README.en.md)*
4+
5+Un agent qui répond **uniquement en JSON** : nommez une pizza, recevez ses ingrédients pour une pizza de 30 cm. La forme de la réponse est déclarée en types Go sous `structuredOutput` dans `agent.yaml`.
6+
7+Modèle : **Gemma 4 E4B** (`hf.co/unsloth/gemma-4-E4B-it-GGUF:Q4_K_M`) sur **llmman**.
8+
9+## Lancer
10+
11+```bash
12+llmman pull hf.co/unsloth/gemma-4-E4B-it-GGUF:Q4_K_M
13+cd demos/structured-output-pizza
14+mm -p "Une pizza quatre fromages" | jq .
15+```
16+
17+En mode prompt (`-p`), stdout ne contient que le JSON : on peut le passer à `jq`.
18+
19+## La déclaration
20+
21+```go
22+// One ingredient of the pizza.
23+type Ingredient struct {
24+ Name string `json:"name"` // the ingredient's name, in the user's language
25+ Quantity float64 `json:"quantity"` // amount for one pizza
26+ Unit string `json:"unit"` // "g", "ml", "piece", "pinch" or "to taste"
27+ Category string `json:"category"` // "base", "sauce", "cheese", "meat", "vegetable", "herb" or "other"
28+ Allergens []string `json:"allergens"` // e.g. "gluten", "lactose"; empty when none
29+ Optional bool `json:"optional"`
30+}
31+
32+type Pizza struct {
33+ Name string `json:"name"`
34+ Vegetarian bool `json:"vegetarian"`
35+ Ingredients []Ingredient `json:"ingredients"`
36+ BakingTime *int `json:"bakingTime"` // minutes, null when unknown
37+}
38+
39+var Recipe Pizza
40+```
41+
42+- Le `var` désigne la racine de la réponse : ici un objet `Pizza`. `var Recipes = []Pizza{}` demanderait un tableau.
43+- Les commentaires `//` deviennent la `description` des champs dans le JSON Schema ; c'est là qu'on liste les valeurs permises, puisque `const` (et donc les énumérations) est refusé.
44+- Toutes les clés sont obligatoires ; un pointeur (`*int`) autorise en plus `null`.
45+- Un agent avec `structuredOutput` ne déclare aucun outil : `bashTool`, `editTools`, `planTool`, `fetchContentTool`, `askUserTool` valent `false` et `skillsDir` vaut `no-skills`.
46+
47+## Observé le 2026-10-16
48+
49+`mm -p "Une pizza quatre fromages"` (environ 40 s) : `"name": "Quatre Fromages"`, `"vegetarian": true`, une pâte, une sauce, quatre fromages (dont `{"name":"Brie","quantity":50.0,"unit":"g","category":"cheese","allergens":["lactose"],"optional":false}`) et du thym facultatif (`"unit":"pinch"`, `"optional":true`).
50+
51+`mm -p "Une pizza margherita"` : cinq ingrédients et `"bakingTime": 12`, avec des noms en italien (`Mozzarella di bufala`, `Basilico fresco`…) — le modèle a préféré la langue du plat à celle de la question.
52+
53+## Pour aller plus loin
54+
55+- Chaque construction Go et son JSON Schema : [référence de configuration — `structuredOutput`](../../docs/fr/reference/configuration.md#structuredoutput)
56+- Pas à pas : [Obtenir des réponses en JSON](../../docs/fr/how-to/get-json-answers.md)
57+- Plusieurs agents JSON et un coordinateur : `demos/structured-output/`
new file mode 100644
@@ -0,0 +1,57 @@
1+# Des réponses JSON à partir de types Go — les ingrédients d'une pizza
2+
3+*[English version](README.en.md)*
4+
5+Un agent qui répond **uniquement en JSON** : nommez une pizza, recevez ses ingrédients pour une pizza de 30 cm. La forme de la réponse est déclarée en types Go sous `structuredOutput` dans `agent.yaml`.
6+
7+Modèle : **Gemma 4 E4B** (`hf.co/unsloth/gemma-4-E4B-it-GGUF:Q4_K_M`) sur **llmman**.
8+
9+## Lancer
10+
11+```bash
12+llmman pull hf.co/unsloth/gemma-4-E4B-it-GGUF:Q4_K_M
13+cd demos/structured-output-pizza
14+mm -p "Une pizza quatre fromages" | jq .
15+```
16+
17+En mode prompt (`-p`), stdout ne contient que le JSON : on peut le passer à `jq`.
18+
19+## La déclaration
20+
21+```go
22+// One ingredient of the pizza.
23+type Ingredient struct {
24+ Name string `json:"name"` // the ingredient's name, in the user's language
25+ Quantity float64 `json:"quantity"` // amount for one pizza
26+ Unit string `json:"unit"` // "g", "ml", "piece", "pinch" or "to taste"
27+ Category string `json:"category"` // "base", "sauce", "cheese", "meat", "vegetable", "herb" or "other"
28+ Allergens []string `json:"allergens"` // e.g. "gluten", "lactose"; empty when none
29+ Optional bool `json:"optional"`
30+}
31+
32+type Pizza struct {
33+ Name string `json:"name"`
34+ Vegetarian bool `json:"vegetarian"`
35+ Ingredients []Ingredient `json:"ingredients"`
36+ BakingTime *int `json:"bakingTime"` // minutes, null when unknown
37+}
38+
39+var Recipe Pizza
40+```
41+
42+- Le `var` désigne la racine de la réponse : ici un objet `Pizza`. `var Recipes = []Pizza{}` demanderait un tableau.
43+- Les commentaires `//` deviennent la `description` des champs dans le JSON Schema ; c'est là qu'on liste les valeurs permises, puisque `const` (et donc les énumérations) est refusé.
44+- Toutes les clés sont obligatoires ; un pointeur (`*int`) autorise en plus `null`.
45+- Un agent avec `structuredOutput` ne déclare aucun outil : `bashTool`, `editTools`, `planTool`, `fetchContentTool`, `askUserTool` valent `false` et `skillsDir` vaut `no-skills`.
46+
47+## Observé le 2026-10-16
48+
49+`mm -p "Une pizza quatre fromages"` (environ 40 s) : `"name": "Quatre Fromages"`, `"vegetarian": true`, une pâte, une sauce, quatre fromages (dont `{"name":"Brie","quantity":50.0,"unit":"g","category":"cheese","allergens":["lactose"],"optional":false}`) et du thym facultatif (`"unit":"pinch"`, `"optional":true`).
50+
51+`mm -p "Une pizza margherita"` : cinq ingrédients et `"bakingTime": 12`, avec des noms en italien (`Mozzarella di bufala`, `Basilico fresco`…) — le modèle a préféré la langue du plat à celle de la question.
52+
53+## Pour aller plus loin
54+
55+- Chaque construction Go et son JSON Schema : [référence de configuration — `structuredOutput`](../../docs/fr/reference/configuration.md#structuredoutput)
56+- Pas à pas : [Obtenir des réponses en JSON](../../docs/fr/how-to/get-json-answers.md)
57+- Plusieurs agents JSON et un coordinateur : `demos/structured-output/`
added demos/structured-output-pizza/agent.yaml +43 -0
new file mode 100644
@@ -0,0 +1,43 @@
1+# Agent settings — "pizza": an agent that answers ONLY in JSON — the
2+# ingredients of the pizza you name, for one 30 cm pizza — with the shape
3+# declared below in `structuredOutput`. mini-me on Gemma 4 E4B, served by
4+# llmman (https://github.com/llmmanorg/llmman).
5+#
6+# llmman pull hf.co/unsloth/gemma-4-E4B-it-GGUF:Q4_K_M
7+# cd demos/structured-output-pizza
8+# mm -p "Une pizza quatre fromages" | jq .
9+provider: llmman
10+model: hf.co/unsloth/gemma-4-E4B-it-GGUF:Q4_K_M
11+baseUrl: http://127.0.0.1:17434/v1
12+fallback: http://host.docker.internal:17434/v1
13+
14+bashTool: false
15+editTools: false
16+planTool: false
17+fetchContentTool: false
18+askUserTool: false
19+skillsDir: no-skills
20+
21+system: |
22+ You are a pizzaiolo. For the pizza the user names, list its ingredients
23+ for one 30 cm pizza.
24+
25+structuredOutput: |
26+ // One ingredient of the pizza.
27+ type Ingredient struct {
28+ Name string `json:"name"` // the ingredient's name, in the user's language
29+ Quantity float64 `json:"quantity"` // amount for one pizza
30+ Unit string `json:"unit"` // "g", "ml", "piece", "pinch" or "to taste"
31+ Category string `json:"category"` // "base", "sauce", "cheese", "meat", "vegetable", "herb" or "other"
32+ Allergens []string `json:"allergens"` // e.g. "gluten", "lactose"; empty when none
33+ Optional bool `json:"optional"`
34+ }
35+
36+ type Pizza struct {
37+ Name string `json:"name"`
38+ Vegetarian bool `json:"vegetarian"`
39+ Ingredients []Ingredient `json:"ingredients"`
40+ BakingTime *int `json:"bakingTime"` // minutes, null when unknown
41+ }
42+
43+ var Recipe Pizza
new file mode 100644
@@ -0,0 +1,43 @@
1+# Agent settings — "pizza": an agent that answers ONLY in JSON — the
2+# ingredients of the pizza you name, for one 30 cm pizza — with the shape
3+# declared below in `structuredOutput`. mini-me on Gemma 4 E4B, served by
4+# llmman (https://github.com/llmmanorg/llmman).
5+#
6+# llmman pull hf.co/unsloth/gemma-4-E4B-it-GGUF:Q4_K_M
7+# cd demos/structured-output-pizza
8+# mm -p "Une pizza quatre fromages" | jq .
9+provider: llmman
10+model: hf.co/unsloth/gemma-4-E4B-it-GGUF:Q4_K_M
11+baseUrl: http://127.0.0.1:17434/v1
12+fallback: http://host.docker.internal:17434/v1
13+
14+bashTool: false
15+editTools: false
16+planTool: false
17+fetchContentTool: false
18+askUserTool: false
19+skillsDir: no-skills
20+
21+system: |
22+ You are a pizzaiolo. For the pizza the user names, list its ingredients
23+ for one 30 cm pizza.
24+
25+structuredOutput: |
26+ // One ingredient of the pizza.
27+ type Ingredient struct {
28+ Name string `json:"name"` // the ingredient's name, in the user's language
29+ Quantity float64 `json:"quantity"` // amount for one pizza
30+ Unit string `json:"unit"` // "g", "ml", "piece", "pinch" or "to taste"
31+ Category string `json:"category"` // "base", "sauce", "cheese", "meat", "vegetable", "herb" or "other"
32+ Allergens []string `json:"allergens"` // e.g. "gluten", "lactose"; empty when none
33+ Optional bool `json:"optional"`
34+ }
35+
36+ type Pizza struct {
37+ Name string `json:"name"`
38+ Vegetarian bool `json:"vegetarian"`
39+ Ingredients []Ingredient `json:"ingredients"`
40+ BakingTime *int `json:"bakingTime"` // minutes, null when unknown
41+ }
42+
43+ var Recipe Pizza