feat(session and litellm): add session deletion flow with Ctrl+P and overlay dialog; support LiteLLM (or any other) proxy; extend vertex models

.gitignore

@@ -44,4 +44,5 @@ Thumbs.db
 .opencode/
 
 opencode
+.aider*
 opencode.md

README.md

@@ -269,6 +269,8 @@ OpenCode supports a variety of AI models from different providers:
 
 - Gemini 2.5
 - Gemini 2.5 Flash
+- Anthropic Sonnet 4
+- Anthropic Opus 4
 
 ## Usage
 
@@ -335,6 +337,7 @@ The output format is implemented as a strongly-typed `OutputFormat` in the codeb
 | `?`      | Toggle help dialog (when not in editing mode)           |
 | `Ctrl+L` | View logs                                               |
 | `Ctrl+A` | Switch session                                          |
+| `Ctrl+P` | Prune session                                           | 
 | `Ctrl+K` | Command dialog                                          |
 | `Ctrl+O` | Toggle model selection dialog                           |
 | `Esc`    | Close current overlay/dialog or return to previous mode |
@@ -626,11 +629,30 @@ This is useful for developers who want to experiment with custom models.
 
 ### Configuring a self-hosted provider
 
-You can use a self-hosted model by setting the `LOCAL_ENDPOINT` environment variable.
-This will cause OpenCode to load and use the models from the specified endpoint.
+You can use a self-hosted model by setting the `LOCAL_ENDPOINT` and `LOCAL_ENDPOINT_API_KEY` environment variable.
+This will cause OpenCode to load and use the models from the specified endpoint. When it loads model it tries to inherit settings
+from predefined ones if possible.
 
 ```bash
 LOCAL_ENDPOINT=http://localhost:1235/v1
+LOCAL_ENDPOINT_API_KEY=secret
+```
+
+### Using LiteLLM Proxy
+It is possible to use LiteLLM as a passthrough proxy by providing `baseURL` and auth header to provider configuration:
+```json
+{
+  "providers": {
+    "vertexai": {
+      "apiKey": "litellm-api-key",
+      "disabled": false,
+      "baseURL": "https://localhost/vertex_ai"
+      "headers": {
+        "x-litellm-api-key": "litellm-api-key"
+      }
+    }
+  }
+}
 ```
 
 ### Configuring a self-hosted model

cmd/schema/main.go

@@ -186,6 +186,17 @@ func generateSchema() map[string]any {
 					"description": "Whether the provider is disabled",
 					"default":     false,
 				},
+				"baseURL": map[string]any{
+					"type":        "string",
+					"description": "Base URL for the provider instead of default one",
+				},
+				"headers": map[string]any{
+					"type":        "object",
+					"description": "Extra headers to attach to request",
+					"additionalProperties": map[string]any{
+						"type": "string",
+					},
+				},
 			},
 		},
 	}

go.mod

@@ -33,9 +33,17 @@ require (
 	github.com/stretchr/testify v1.10.0
 )
 
+require (
+	cloud.google.com/go/auth/oauth2adapt v0.2.6 // indirect
+	go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.54.0 // indirect
+	golang.org/x/oauth2 v0.25.0 // indirect
+	golang.org/x/time v0.8.0 // indirect
+	google.golang.org/api v0.215.0 // indirect
+)
+
 require (
 	cloud.google.com/go v0.116.0 // indirect
-	cloud.google.com/go/auth v0.13.0 // indirect
+	cloud.google.com/go/auth v0.13.0
 	cloud.google.com/go/compute/metadata v0.6.0 // indirect
 	github.com/Azure/azure-sdk-for-go/sdk/azcore v1.17.0 // indirect
 	github.com/Azure/azure-sdk-for-go/sdk/internal v1.10.0 // indirect

go.sum

@@ -2,6 +2,8 @@ cloud.google.com/go v0.116.0 h1:B3fRrSDkLRt5qSHWe40ERJvhvnQwdZiHu0bJOpldweE=
 cloud.google.com/go v0.116.0/go.mod h1:cEPSRWPzZEswwdr9BxE6ChEn01dWlTaF05LiC2Xs70U=
 cloud.google.com/go/auth v0.13.0 h1:8Fu8TZy167JkW8Tj3q7dIkr2v4cndv41ouecJx0PAHs=
 cloud.google.com/go/auth v0.13.0/go.mod h1:COOjD9gwfKNKz+IIduatIhYJQIc0mG3H102r/EMxX6Q=
+cloud.google.com/go/auth/oauth2adapt v0.2.6 h1:V6a6XDu2lTwPZWOawrAa9HUK+DB2zfJyTuciBG5hFkU=
+cloud.google.com/go/auth/oauth2adapt v0.2.6/go.mod h1:AlmsELtlEBnaNTL7jCj8VQFLy6mbZv0s4Q7NGBeQ5E8=
 cloud.google.com/go/compute/metadata v0.6.0 h1:A6hENjEsCDtC1k8byVsgwvVcioamEHvZ4j01OwKxG9I=
 cloud.google.com/go/compute/metadata v0.6.0/go.mod h1:FjyFAW1MW0C203CEOMDTu3Dk1FlqW3Rga40jzHL4hfg=
 github.com/Azure/azure-sdk-for-go/sdk/azcore v1.17.0 h1:g0EZJwz7xkXQiZAI5xi9f3WWFYBlX1CPTrR+NDToRkQ=
@@ -250,6 +252,8 @@ github.com/yuin/goldmark-emoji v1.0.5 h1:EMVWyCGPlXJfUXBXpuMu+ii3TIaxbVBnEX9uaDC
 github.com/yuin/goldmark-emoji v1.0.5/go.mod h1:tTkZEbwu5wkPmgTcitqddVxY9osFZiavD+r4AzQrh1U=
 go.opentelemetry.io/auto/sdk v1.1.0 h1:cH53jehLUN6UFLY71z+NDOiNJqDdPRaXzTel0sJySYA=
 go.opentelemetry.io/auto/sdk v1.1.0/go.mod h1:3wSPjt5PWp2RhlCcmmOial7AvC4DQqZb7a7wCow3W8A=
+go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.54.0 h1:r6I7RJCN86bpD/FQwedZ0vSixDpwuWREjW9oRMsmqDc=
+go.opentelemetry.io/contrib/instrumentation/google.golang.org/grpc/otelgrpc v0.54.0/go.mod h1:B9yO6b04uB80CzjedvewuqDhxJxi11s7/GtiGa8bAjI=
 go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.54.0 h1:TT4fX+nBOA/+LUkobKGW1ydGcn+G3vRw9+g5HwCphpk=
 go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.54.0/go.mod h1:L7UH0GbB0p47T4Rri3uHjbpCFYrVrwc1I25QhNPiGK8=
 go.opentelemetry.io/otel v1.35.0 h1:xKWKPxrxB6OtMCbmMY021CqC45J+3Onta9MqjhnusiQ=
@@ -289,6 +293,8 @@ golang.org/x/net v0.24.0/go.mod h1:2Q7sJY5mzlzWjKtYUEXSlBWCdyaioyXzRB2RtU8KVE8=
 golang.org/x/net v0.25.0/go.mod h1:JkAGAh7GEvH74S6FOH42FLoXpXbE/aqXSrIQjXgsiwM=
 golang.org/x/net v0.39.0 h1:ZCu7HMWDxpXpaiKdhzIfaltL9Lp31x/3fCP11bc6/fY=
 golang.org/x/net v0.39.0/go.mod h1:X7NRbYVEA+ewNkCNyJ513WmMdQ3BineSwVtN2zD/d+E=
+golang.org/x/oauth2 v0.25.0 h1:CY4y7XT9v0cRI9oupztF8AgiIu99L/ksR/Xp/6jrZ70=
+golang.org/x/oauth2 v0.25.0/go.mod h1:XYTD2NtWslqkgxebSiOHnXEap4TF09sJSc7H1sXbhtI=
 golang.org/x/sync v0.0.0-20190423024810-112230192c58/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.0.0-20220722155255-886fb9371eb4/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
 golang.org/x/sync v0.1.0/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
@@ -329,11 +335,15 @@ golang.org/x/text v0.14.0/go.mod h1:18ZOQIKpY8NJVqYksKHtTdi31H5itFRjB5/qKTNYzSU=
 golang.org/x/text v0.15.0/go.mod h1:18ZOQIKpY8NJVqYksKHtTdi31H5itFRjB5/qKTNYzSU=
 golang.org/x/text v0.24.0 h1:dd5Bzh4yt5KYA8f9CJHCP4FB4D51c2c6JvN37xJJkJ0=
 golang.org/x/text v0.24.0/go.mod h1:L8rBsPeo2pSS+xqN0d5u2ikmjtmoJbDBT1b7nHvFCdU=
+golang.org/x/time v0.8.0 h1:9i3RxcPv3PZnitoVGMPDKZSq1xW1gK1Xy3ArNOGZfEg=
+golang.org/x/time v0.8.0/go.mod h1:3BpzKBy/shNhVucY/MWOyx10tF3SFh9QdLuxbVysPQM=
 golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
 golang.org/x/tools v0.0.0-20191119224855-298f0cb1881e/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
 golang.org/x/tools v0.1.12/go.mod h1:hNGJHUnrk76NpqgfD5Aqm5Crs+Hm0VOH/i9J2+nxYbc=
 golang.org/x/tools v0.6.0/go.mod h1:Xwgl3UAJ/d3gWutnCtw505GrjyAbvKui8lOU390QaIU=
 golang.org/x/xerrors v0.0.0-20190717185122-a985d3407aa7/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
+google.golang.org/api v0.215.0 h1:jdYF4qnyczlEz2ReWIsosNLDuzXyvFHJtI5gcr0J7t0=
+google.golang.org/api v0.215.0/go.mod h1:fta3CVtuJYOEdugLNWm6WodzOS8KdFckABwN4I40hzY=
 google.golang.org/genai v1.3.0 h1:tXhPJF30skOjnnDY7ZnjK3q7IKy4PuAlEA0fk7uEaEI=
 google.golang.org/genai v1.3.0/go.mod h1:TyfOKRz/QyCaj6f/ZDt505x+YreXnY40l2I6k8TvgqY=
 google.golang.org/genproto/googleapis/rpc v0.0.0-20250324211829-b45e905df463 h1:e0AIkUUhxyBKh6ssZNrAMeqhA7RKUj42346d1y02i2g=

internal/config/config.go

@@ -52,8 +52,10 @@ type Agent struct {
 
 // Provider defines configuration for an LLM provider.
 type Provider struct {
-	APIKey   string `json:"apiKey"`
-	Disabled bool   `json:"disabled"`
+	APIKey   string            `json:"apiKey"`
+	Disabled bool              `json:"disabled"`
+	BaseURL  string            `json:"baseURL"`
+	Headers  map[string]string `json:"headers,omitempty"`
 }
 
 // Data defines storage configuration.

internal/llm/agent/agent.go

@@ -725,12 +725,20 @@ func createAgentProvider(agentName config.AgentName) (provider.Provider, error)
 	if agentConfig.MaxTokens > 0 {
 		maxTokens = agentConfig.MaxTokens
 	}
+
 	opts := []provider.ProviderClientOption{
 		provider.WithAPIKey(providerCfg.APIKey),
 		provider.WithModel(model),
 		provider.WithSystemMessage(prompt.GetAgentPrompt(agentName, model.Provider)),
 		provider.WithMaxTokens(maxTokens),
 	}
+	if providerCfg.BaseURL != "" {
+		opts = append(opts, provider.WithBaseURL(providerCfg.BaseURL))
+	}
+	if len(providerCfg.Headers) != 0 {
+		opts = append(opts, provider.WithHeaders(providerCfg.Headers))
+	}
+
 	if model.Provider == models.ProviderOpenAI || model.Provider == models.ProviderLocal && model.CanReason {
 		opts = append(
 			opts,

internal/llm/models/anthropic.go

@@ -91,7 +91,7 @@ var AnthropicModels = map[ModelID]Model{
 		CostPer1MOutCached:  0.30,
 		CostPer1MOut:        15.0,
 		ContextWindow:       200000,
-		DefaultMaxTokens:    50000,
+		DefaultMaxTokens:    64000,
 		CanReason:           true,
 		SupportsAttachments: true,
 	},
@@ -105,7 +105,7 @@ var AnthropicModels = map[ModelID]Model{
 		CostPer1MOutCached:  1.50,
 		CostPer1MOut:        75.0,
 		ContextWindow:       200000,
-		DefaultMaxTokens:    4096,
+		DefaultMaxTokens:    32000,
 		SupportsAttachments: true,
 	},
 }

internal/llm/models/bootstrap.go

@@ -0,0 +1,18 @@
+package models
+
+import "maps"
+
+func init() {
+	maps.Copy(SupportedModels, AnthropicModels)
+	maps.Copy(SupportedModels, OpenAIModels)
+	maps.Copy(SupportedModels, GeminiModels)
+	maps.Copy(SupportedModels, GroqModels)
+	maps.Copy(SupportedModels, AzureModels)
+	maps.Copy(SupportedModels, OpenRouterModels)
+	maps.Copy(SupportedModels, XAIModels)
+	maps.Copy(SupportedModels, VertexAIGeminiModels)
+	maps.Copy(SupportedModels, VertexAIAnthropicModels)
+	maps.Copy(SupportedModels, CopilotModels)
+
+	initLocalModels()
+}

internal/llm/models/local.go

@@ -3,6 +3,7 @@ package models
 import (
 	"cmp"
 	"encoding/json"
+	"fmt"
 	"net/http"
 	"net/url"
 	"os"
@@ -21,7 +22,7 @@ const (
 	lmStudioBetaModelsPath = "api/v0/models"
 )
 
-func init() {
+func initLocalModels() {
 	if endpoint := os.Getenv("LOCAL_ENDPOINT"); endpoint != "" {
 		localEndpoint, err := url.Parse(endpoint)
 		if err != nil {
@@ -33,7 +34,8 @@ func init() {
 		}
 
 		load := func(url *url.URL, path string) []localModel {
-			url.Path = path
+			url = url.JoinPath(path)
+			logging.Debug(fmt.Sprintf("Trying to load models from %s", url))
 			return listLocalModels(url.String())
 		}
 
@@ -43,16 +45,22 @@ func init() {
 			models = load(localEndpoint, localModelsPath)
 		}
 
-		if len(models) == 0 {
+		if c := len(models); c == 0 {
 			logging.Debug("No local models found",
 				"endpoint", endpoint,
 			)
 			return
+		} else {
+			logging.Debug(fmt.Sprintf("%d local models found", c))
 		}
 
 		loadLocalModels(models)
 
-		viper.SetDefault("providers.local.apiKey", "dummy")
+		if token, ok := os.LookupEnv("LOCAL_ENDPOINT_API_KEY"); ok {
+			viper.SetDefault("providers.local.apiKey", token)
+		} else {
+			viper.SetDefault("providers.local.apiKey", "dummy")
+		}
 		ProviderPopularity[ProviderLocal] = 0
 	}
 }
@@ -75,8 +83,26 @@ type localModel struct {
 }
 
 func listLocalModels(modelsEndpoint string) []localModel {
-	res, err := http.Get(modelsEndpoint)
-	if err != nil {
+	token := os.Getenv("LOCAL_ENDPOINT_API_KEY")
+	var (
+		res *http.Response
+		err error
+	)
+	if token != "" {
+		req, reqErr := http.NewRequest("GET", modelsEndpoint, nil)
+		if reqErr != nil {
+			logging.Debug("Failed to create local models request",
+				"error", reqErr,
+				"endpoint", modelsEndpoint,
+			)
+			return nil
+		}
+		req.Header.Set("Authorization", "Bearer "+token)
+		res, err = http.DefaultClient.Do(req)
+	} else {
+		res, err = http.Get(modelsEndpoint)
+	}
+	if err != nil || res == nil {
 		logging.Debug("Failed to list local models",
 			"error", err,
 			"endpoint", modelsEndpoint,
@@ -125,7 +151,8 @@ func listLocalModels(modelsEndpoint string) []localModel {
 
 func loadLocalModels(models []localModel) {
 	for i, m := range models {
-		model := convertLocalModel(m)
+		source := tryResolveSource(m.ID)
+		model := convertLocalModel(m, source)
 		SupportedModels[model.ID] = model
 
 		if i == 0 || m.State == "loaded" {
@@ -137,16 +164,34 @@ func loadLocalModels(models []localModel) {
 	}
 }
 
-func convertLocalModel(model localModel) Model {
-	return Model{
-		ID:                  ModelID("local." + model.ID),
-		Name:                friendlyModelName(model.ID),
-		Provider:            ProviderLocal,
-		APIModel:            model.ID,
-		ContextWindow:       cmp.Or(model.LoadedContextLength, 4096),
-		DefaultMaxTokens:    cmp.Or(model.LoadedContextLength, 4096),
-		CanReason:           true,
-		SupportsAttachments: true,
+func tryResolveSource(localID string) *Model {
+	for _, model := range SupportedModels {
+		if strings.Contains(localID, model.APIModel) {
+			return &model
+		}
+	}
+	return nil
+}
+
+func convertLocalModel(model localModel, source *Model) Model {
+	if source != nil {
+		m := *source
+		m.ID = ModelID("local." + model.ID)
+		m.Name = source.Name
+		m.APIModel = model.ID
+		m.Provider = ProviderLocal
+		return m
+	} else {
+		return Model{
+			ID:                  ModelID("local." + model.ID),
+			Name:                friendlyModelName(model.ID),
+			Provider:            ProviderLocal,
+			APIModel:            model.ID,
+			ContextWindow:       cmp.Or(model.LoadedContextLength, 4096),
+			DefaultMaxTokens:    cmp.Or(model.LoadedContextLength, 4096),
+			CanReason:           false,
+			SupportsAttachments: false,
+		}
 	}
 }
 

internal/llm/models/vertexai.go

@@ -6,6 +6,8 @@ const (
 	// Models
 	VertexAIGemini25Flash ModelID = "vertexai.gemini-2.5-flash"
 	VertexAIGemini25      ModelID = "vertexai.gemini-2.5"
+	VertexAISonnet4       ModelID = "vertexai.claude-sonnet-4"
+	VertexAIOpus4         ModelID = "vertexai.claude-opus-4"
 )
 
 var VertexAIGeminiModels = map[ModelID]Model{
@@ -36,3 +38,34 @@ var VertexAIGeminiModels = map[ModelID]Model{
 		SupportsAttachments: true,
 	},
 }
+
+var VertexAIAnthropicModels = map[ModelID]Model{
+	VertexAISonnet4: {
+		ID:                  VertexAISonnet4,
+		Name:                "VertexAI: Claude Sonnet 4",
+		Provider:            ProviderVertexAI,
+		APIModel:            "claude-sonnet-4",
+		CostPer1MIn:         AnthropicModels[Claude4Sonnet].CostPer1MIn,
+		CostPer1MInCached:   AnthropicModels[Claude4Sonnet].CostPer1MInCached,
+		CostPer1MOut:        AnthropicModels[Claude4Sonnet].CostPer1MOut,
+		CostPer1MOutCached:  AnthropicModels[Claude4Sonnet].CostPer1MOutCached,
+		ContextWindow:       AnthropicModels[Claude4Sonnet].ContextWindow,
+		DefaultMaxTokens:    AnthropicModels[Claude4Sonnet].DefaultMaxTokens,
+		SupportsAttachments: AnthropicModels[Claude4Sonnet].SupportsAttachments,
+		CanReason:           AnthropicModels[Claude4Sonnet].CanReason,
+	},
+	VertexAIOpus4: {
+		ID:                  VertexAIOpus4,
+		Name:                "VertexAI: Claude Opus 4",
+		Provider:            ProviderVertexAI,
+		APIModel:            "claude-opus-4@20250514",
+		CostPer1MIn:         AnthropicModels[Claude4Opus].CostPer1MIn,
+		CostPer1MInCached:   AnthropicModels[Claude4Opus].CostPer1MInCached,
+		CostPer1MOut:        AnthropicModels[Claude4Opus].CostPer1MOut,
+		CostPer1MOutCached:  AnthropicModels[Claude4Opus].CostPer1MOutCached,
+		ContextWindow:       AnthropicModels[Claude4Opus].ContextWindow,
+		DefaultMaxTokens:    AnthropicModels[Claude4Opus].DefaultMaxTokens,
+		SupportsAttachments: AnthropicModels[Claude4Opus].SupportsAttachments,
+		CanReason:           AnthropicModels[Claude4Opus].CanReason,
+	},
+}