feat(ai): add support for Ollama AI provider configuration

* Update README with Ollama integration details
* Add Ollama configuration to example YAML files
* Implement Ollama provider in AI factory
* Add tests for Ollama provider functionality
* Enhance config validation for Ollama settings

README.md

@@ -16,8 +16,9 @@ A Go MCP server for capturing and retrieving thoughts, memory, and project conte
 
 - Go — MCP server over Streamable HTTP
 - Postgres + pgvector — storage and vector search
-- LiteLLM — primary AI provider (embeddings + metadata extraction)
+- LiteLLM — primary hosted AI provider (embeddings + metadata extraction)
 - OpenRouter — default upstream behind LiteLLM
+- Ollama — supported local or self-hosted OpenAI-compatible provider
 
 ## Tools
 
@@ -48,6 +49,7 @@ Config is YAML-driven. Copy `configs/config.example.yaml` and set:
 - `database.url` — Postgres connection string
 - `auth.keys` — API keys for MCP endpoint access
 - `ai.litellm.base_url` and `ai.litellm.api_key` — LiteLLM proxy
+- `ai.ollama.base_url` and `ai.ollama.api_key` — Ollama local or remote server
 
 See `llm/plan.md` for full architecture and implementation plan.
 
@@ -72,7 +74,35 @@ Notes:
 
 - The app uses `configs/docker.yaml` inside the container.
 - `OB1_LITELLM_BASE_URL` overrides the LiteLLM endpoint, so you can retarget it without editing YAML.
+- `OB1_OLLAMA_BASE_URL` overrides the Ollama endpoint for local or remote servers.
 - The base Compose file uses `host.containers.internal`, which is Podman-friendly.
 - The Docker override file adds `host-gateway` aliases so Docker can resolve the same host endpoint.
 - Database migrations `001` through `005` run automatically when the Postgres volume is created for the first time.
 - `migrations/006_rls_and_grants.sql` is intentionally skipped during container bootstrap because it contains deployment-specific grants for a role named `amcs_user`.
+
+## Ollama
+
+Set `ai.provider: "ollama"` to use a local or self-hosted Ollama server through its OpenAI-compatible API.
+
+Example:
+
+```yaml
+ai:
+  provider: "ollama"
+  embeddings:
+    model: "nomic-embed-text"
+    dimensions: 768
+  metadata:
+    model: "llama3.2"
+    temperature: 0.1
+  ollama:
+    base_url: "http://localhost:11434/v1"
+    api_key: "ollama"
+    request_headers: {}
+```
+
+Notes:
+
+- For remote Ollama servers, point `ai.ollama.base_url` at the remote `/v1` endpoint.
+- The client always sends Bearer auth; Ollama ignores it locally, so `api_key: "ollama"` is a safe default.
+- `ai.embeddings.dimensions` must match the embedding model you actually use, or startup will fail the database vector-dimension check.

configs/config.example.yaml

@@ -45,6 +45,10 @@ ai:
     request_headers: {}
     embedding_model: "openrouter/openai/text-embedding-3-small"
     metadata_model: "gpt-4o-mini"
+  ollama:
+    base_url: "http://localhost:11434/v1"
+    api_key: "ollama"
+    request_headers: {}
   openrouter:
     base_url: "https://openrouter.ai/api/v1"
     api_key: ""

configs/dev.yaml

@@ -45,6 +45,10 @@ ai:
     request_headers: {}
     embedding_model: "openrouter/openai/text-embedding-3-small"
     metadata_model: "gpt-4o-mini"
+  ollama:
+    base_url: "http://localhost:11434/v1"
+    api_key: "ollama"
+    request_headers: {}
   openrouter:
     base_url: "https://openrouter.ai/api/v1"
     api_key: ""

configs/docker.yaml

@@ -45,6 +45,10 @@ ai:
     request_headers: {}
     embedding_model: "openrouter/openai/text-embedding-3-small"
     metadata_model: "gpt-4o-mini"
+  ollama:
+    base_url: "http://host.containers.internal:11434/v1"
+    api_key: "ollama"
+    request_headers: {}
   openrouter:
     base_url: "https://openrouter.ai/api/v1"
     api_key: ""

internal/ai/factory.go

@@ -6,6 +6,7 @@ import (
 	"net/http"
 
 	"git.warky.dev/wdevs/amcs/internal/ai/litellm"
+	"git.warky.dev/wdevs/amcs/internal/ai/ollama"
 	"git.warky.dev/wdevs/amcs/internal/ai/openrouter"
 	"git.warky.dev/wdevs/amcs/internal/config"
 )
@@ -14,6 +15,8 @@ func NewProvider(cfg config.AIConfig, httpClient *http.Client, log *slog.Logger)
 	switch cfg.Provider {
 	case "litellm":
 		return litellm.New(cfg, httpClient, log)
+	case "ollama":
+		return ollama.New(cfg, httpClient, log)
 	case "openrouter":
 		return openrouter.New(cfg, httpClient, log)
 	default:

internal/ai/factory_test.go

@@ -0,0 +1,33 @@
+package ai
+
+import (
+	"io"
+	"log/slog"
+	"net/http"
+	"testing"
+
+	"git.warky.dev/wdevs/amcs/internal/config"
+)
+
+func TestNewProviderSupportsOllama(t *testing.T) {
+	provider, err := NewProvider(config.AIConfig{
+		Provider: "ollama",
+		Embeddings: config.AIEmbeddingConfig{
+			Model:      "nomic-embed-text",
+			Dimensions: 768,
+		},
+		Metadata: config.AIMetadataConfig{
+			Model: "llama3.2",
+		},
+		Ollama: config.OllamaConfig{
+			BaseURL: "http://localhost:11434/v1",
+			APIKey:  "ollama",
+		},
+	}, &http.Client{}, slog.New(slog.NewTextHandler(io.Discard, nil)))
+	if err != nil {
+		t.Fatalf("NewProvider() error = %v", err)
+	}
+	if provider.Name() != "ollama" {
+		t.Fatalf("provider name = %q, want ollama", provider.Name())
+	}
+}

internal/ai/ollama/client.go

@@ -0,0 +1,24 @@
+package ollama
+
+import (
+	"log/slog"
+	"net/http"
+
+	"git.warky.dev/wdevs/amcs/internal/ai/compat"
+	"git.warky.dev/wdevs/amcs/internal/config"
+)
+
+func New(cfg config.AIConfig, httpClient *http.Client, log *slog.Logger) (*compat.Client, error) {
+	return compat.New(compat.Config{
+		Name:           "ollama",
+		BaseURL:        cfg.Ollama.BaseURL,
+		APIKey:         cfg.Ollama.APIKey,
+		EmbeddingModel: cfg.Embeddings.Model,
+		MetadataModel:  cfg.Metadata.Model,
+		Temperature:    cfg.Metadata.Temperature,
+		Headers:        cfg.Ollama.RequestHeaders,
+		HTTPClient:     httpClient,
+		Log:            log,
+		Dimensions:     cfg.Embeddings.Dimensions,
+	}), nil
+}

internal/config/config.go

@@ -62,6 +62,7 @@ type AIConfig struct {
 	Embeddings AIEmbeddingConfig  `yaml:"embeddings"`
 	Metadata   AIMetadataConfig   `yaml:"metadata"`
 	LiteLLM    LiteLLMConfig      `yaml:"litellm"`
+	Ollama     OllamaConfig       `yaml:"ollama"`
 	OpenRouter OpenRouterAIConfig `yaml:"openrouter"`
 }
 
@@ -84,6 +85,12 @@ type LiteLLMConfig struct {
 	MetadataModel   string            `yaml:"metadata_model"`
 }
 
+type OllamaConfig struct {
+	BaseURL        string            `yaml:"base_url"`
+	APIKey         string            `yaml:"api_key"`
+	RequestHeaders map[string]string `yaml:"request_headers"`
+}
+
 type OpenRouterAIConfig struct {
 	BaseURL      string            `yaml:"base_url"`
 	APIKey       string            `yaml:"api_key"`

internal/config/loader.go

@@ -73,6 +73,10 @@ func defaultConfig() Config {
 				Model:       "gpt-4o-mini",
 				Temperature: 0.1,
 			},
+			Ollama: OllamaConfig{
+				BaseURL: "http://localhost:11434/v1",
+				APIKey:  "ollama",
+			},
 		},
 		Capture: CaptureConfig{
 			Source: DefaultSource,
@@ -97,6 +101,8 @@ func applyEnvOverrides(cfg *Config) {
 	overrideString(&cfg.Database.URL, "OB1_DATABASE_URL")
 	overrideString(&cfg.AI.LiteLLM.BaseURL, "OB1_LITELLM_BASE_URL")
 	overrideString(&cfg.AI.LiteLLM.APIKey, "OB1_LITELLM_API_KEY")
+	overrideString(&cfg.AI.Ollama.BaseURL, "OB1_OLLAMA_BASE_URL")
+	overrideString(&cfg.AI.Ollama.APIKey, "OB1_OLLAMA_API_KEY")
 	overrideString(&cfg.AI.OpenRouter.APIKey, "OB1_OPENROUTER_API_KEY")
 
 	if value, ok := os.LookupEnv("OB1_SERVER_PORT"); ok {

internal/config/loader_test.go

@@ -69,3 +69,51 @@ logging:
 		t.Fatalf("server port = %d, want 9090", cfg.Server.Port)
 	}
 }
+
+func TestLoadAppliesOllamaEnvOverrides(t *testing.T) {
+	configPath := filepath.Join(t.TempDir(), "test.yaml")
+	if err := os.WriteFile(configPath, []byte(`
+server:
+  port: 8080
+mcp:
+  path: "/mcp"
+auth:
+  keys:
+    - id: "test"
+      value: "secret"
+database:
+  url: "postgres://from-file"
+ai:
+  provider: "ollama"
+  embeddings:
+    model: "nomic-embed-text"
+    dimensions: 768
+  metadata:
+    model: "llama3.2"
+  ollama:
+    base_url: "http://localhost:11434/v1"
+    api_key: "ollama"
+search:
+  default_limit: 10
+  max_limit: 50
+logging:
+  level: "info"
+`), 0o600); err != nil {
+		t.Fatalf("write config: %v", err)
+	}
+
+	t.Setenv("OB1_OLLAMA_BASE_URL", "https://ollama.example.com/v1")
+	t.Setenv("OB1_OLLAMA_API_KEY", "remote-key")
+
+	cfg, _, err := Load(configPath)
+	if err != nil {
+		t.Fatalf("Load() error = %v", err)
+	}
+
+	if cfg.AI.Ollama.BaseURL != "https://ollama.example.com/v1" {
+		t.Fatalf("ollama base url = %q, want env override", cfg.AI.Ollama.BaseURL)
+	}
+	if cfg.AI.Ollama.APIKey != "remote-key" {
+		t.Fatalf("ollama api key = %q, want env override", cfg.AI.Ollama.APIKey)
+	}
+}

internal/config/validate.go

@@ -28,7 +28,7 @@ func (c Config) Validate() error {
 	}
 
 	switch c.AI.Provider {
-	case "litellm", "openrouter":
+	case "litellm", "ollama", "openrouter":
 	default:
 		return fmt.Errorf("invalid config: unsupported ai.provider %q", c.AI.Provider)
 	}
@@ -45,6 +45,13 @@ func (c Config) Validate() error {
 		if strings.TrimSpace(c.AI.LiteLLM.APIKey) == "" {
 			return fmt.Errorf("invalid config: ai.litellm.api_key is required when ai.provider=litellm")
 		}
+	case "ollama":
+		if strings.TrimSpace(c.AI.Ollama.BaseURL) == "" {
+			return fmt.Errorf("invalid config: ai.ollama.base_url is required when ai.provider=ollama")
+		}
+		if strings.TrimSpace(c.AI.Ollama.APIKey) == "" {
+			return fmt.Errorf("invalid config: ai.ollama.api_key is required when ai.provider=ollama")
+		}
 	case "openrouter":
 		if strings.TrimSpace(c.AI.OpenRouter.BaseURL) == "" {
 			return fmt.Errorf("invalid config: ai.openrouter.base_url is required when ai.provider=openrouter")

internal/config/validate_test.go

@@ -19,6 +19,10 @@ func validConfig() Config {
 				BaseURL: "http://localhost:4000/v1",
 				APIKey:  "key",
 			},
+			Ollama: OllamaConfig{
+				BaseURL: "http://localhost:11434/v1",
+				APIKey:  "ollama",
+			},
 			OpenRouter: OpenRouterAIConfig{
 				BaseURL: "https://openrouter.ai/api/v1",
 				APIKey:  "key",
@@ -29,12 +33,17 @@ func validConfig() Config {
 	}
 }
 
-func TestValidateAcceptsLiteLLMAndOpenRouter(t *testing.T) {
+func TestValidateAcceptsSupportedProviders(t *testing.T) {
 	cfg := validConfig()
 	if err := cfg.Validate(); err != nil {
 		t.Fatalf("Validate litellm error = %v", err)
 	}
 
+	cfg.AI.Provider = "ollama"
+	if err := cfg.Validate(); err != nil {
+		t.Fatalf("Validate ollama error = %v", err)
+	}
+
 	cfg.AI.Provider = "openrouter"
 	if err := cfg.Validate(); err != nil {
 		t.Fatalf("Validate openrouter error = %v", err)