Skip to content

Open source AI coding agent. Designed for large projects and real world tasks.

Add local model support and improve provider handling

Dane Schneidereb0db5f6ca53
diff --git a/app/cli/cmd/model_providers.go b/app/cli/cmd/model_providers.go
index 5781c540..04b646a9 100644
--- a/app/cli/cmd/model_providers.go
+++ b/app/cli/cmd/model_providers.go
@@ -77,7 +77,14 @@ func listProviders(cmd *cobra.Command, args []string) {
color.New(color.Bold, term.ColorHiCyan).Println("🏠 Built-in Providers")
table := tablewriter.NewWriter(os.Stdout)
table.SetAutoWrapText(true)
- table.SetHeader([]string{"ID", "Base URL", "API Key", "Other Vars"})
+
+ var header []string
+ if auth.Current.IsCloud {
+ header = []string{"ID", "API Key", "Other Vars"}
+ } else {
+ header = []string{"ID", "Base URL", "API Key", "Other Vars"}
+ }
+ table.SetHeader(header)
for _, p := range shared.AllModelProviders {
if p == shared.ModelProviderCustom {
continue
@@ -94,16 +101,27 @@ func listProviders(cmd *cobra.Command, args []string) {
}
extraVars := []string{}
+ if config.Provider == shared.ModelProviderAmazonBedrock {
+ extraVars = append(extraVars, "PLANDEX_AWS_PROFILE")
+ }
for _, v := range config.ExtraAuthVars {
extraVars = append(extraVars, v.Var)
}
- table.Append([]string{
- string(p),
- config.BaseUrl,
- apiKey,
- strings.Join(extraVars, "\n"),
- })
+ if auth.Current.IsCloud {
+ table.Append([]string{
+ string(p),
+ apiKey,
+ strings.Join(extraVars, "\n"),
+ })
+ } else {
+ table.Append([]string{
+ string(p),
+ config.BaseUrl,
+ apiKey,
+ strings.Join(extraVars, "\n"),
+ })
+ }
}
table.Render()
fmt.Println()
@@ -135,5 +153,9 @@ func listProviders(cmd *cobra.Command, args []string) {
fmt.Println()
}
+ fmt.Println(color.New(color.Bold, term.ColorHiCyan).Sprint("\n📖 Per-provider instructions"))
+ fmt.Println("Go to → " + color.New(color.Bold).Sprint("https://docs.plandex.ai/models/model-providers"))
+ fmt.Println()
+
term.PrintCmds("", "models custom")
}
diff --git a/app/cli/lib/model_credentials.go b/app/cli/lib/model_credentials.go
index 793036e6..a21d12eb 100644
--- a/app/cli/lib/model_credentials.go
+++ b/app/cli/lib/model_credentials.go
@@ -285,7 +285,7 @@ func mergeAuthVars(dest, src map[string]string) {
func showCredentialErrorMessage(res CredentialCheckResult, opts shared.ModelProviderOptions) {
boldRed := color.New(color.Bold, term.ColorHiRed)
cyanChip := color.New(color.BgCyan, color.FgHiWhite)
- fmt.Println(boldRed.Sprint("🚨 Required model credentials are missing"))
+ fmt.Println(boldRed.Sprint("🚨 Required API key(s) or model credentials are missing"))
someOK, someMissing := false, false
for _, p := range res.Publishers {
@@ -415,7 +415,6 @@ func showCredentialErrorMessage(res CredentialCheckResult, opts shared.ModelProv
fmt.Printf("%s → %s\n", color.New(color.Bold).Sprint(p+" models"), strings.Join(providerNames, ", "))
}
- // TODO: set correct link
fmt.Println(color.New(color.Bold, term.ColorHiCyan).Sprint("\n📖 Per-provider instructions"))
fmt.Println("For details on the API key/credentials required for each provider, go to:\n" + color.New(color.Bold).Sprint("https://docs.plandex.ai/models/model-providers"))
}
diff --git a/app/shared/ai_models_available.go b/app/shared/ai_models_available.go
index 8ad074a6..252120d8 100644
--- a/app/shared/ai_models_available.go
+++ b/app/shared/ai_models_available.go
@@ -297,7 +297,15 @@ var BuiltInModels = []*BaseModelConfigSchema{
},
Variants: []BaseModelConfigVariant{
{IsBaseVariant: true},
- {VariantTag: "thinking", Description: "thinking", Overrides: BaseModelShared{IncludeReasoning: true}},
+ {
+ VariantTag: "thinking",
+ Description: "thinking",
+ Overrides: BaseModelShared{IncludeReasoning: true},
+ Variants: []BaseModelConfigVariant{
+ {VariantTag: "visible", IsDefaultVariant: true, Description: "visible"},
+ {VariantTag: "hidden", Description: "hidden", Overrides: BaseModelShared{HideReasoning: true}},
+ },
+ },
},
Providers: []BaseModelUsesProvider{
{Provider: ModelProviderGoogleAIStudio, ModelName: "gemini/gemini-2.5-flash"},
@@ -389,142 +397,192 @@ var BuiltInModels = []*BaseModelConfigSchema{
{Provider: ModelProviderOllama, ModelName: "ollama_chat/deepseek-r1:8b"},
},
},
+
{
- ModelTag: "perplexity/r1-1776",
- Publisher: ModelPublisherPerplexity,
- Description: "Perplexity R1-1776",
+ ModelTag: "qwen/qwen-2.5-coder-32b-instruct",
+ Publisher: ModelPublisherQwen,
+ Description: "Qwen 2.5 Coder 32B (Instruct)",
BaseModelShared: BaseModelShared{
- DefaultMaxConvoTokens: 7500, MaxTokens: 128000,
- MaxOutputTokens: 128000, ReservedOutputTokens: 30000,
+ DefaultMaxConvoTokens: 10000, MaxTokens: 128000,
+ MaxOutputTokens: 8192, ReservedOutputTokens: 8192,
PreferredOutputFormat: ModelOutputFormatXml,
},
- Variants: []BaseModelConfigVariant{
- {VariantTag: "visible", IsDefaultVariant: true, Description: "(reasoning visible)", Overrides: BaseModelShared{IncludeReasoning: true}},
- {VariantTag: "hidden", Description: "(reasoning hidden)", Overrides: BaseModelShared{IncludeReasoning: false}},
- },
Providers: []BaseModelUsesProvider{
- {Provider: ModelProviderPerplexity, ModelName: "r1-1776-online"},
- {Provider: ModelProviderOpenRouter, ModelName: "perplexity/r1-1776"},
+ {Provider: ModelProviderOpenRouter, ModelName: "qwen/qwen-2.5-coder-32b-instruct"},
},
},
-
{
- ModelTag: "perplexity/r1-1776-70b",
- Publisher: ModelPublisherPerplexity,
- Description: "Perplexity R1-1776 70B",
+ ModelTag: "qwen/qwen3-235b-local",
+ Publisher: ModelPublisherQwen,
+ Description: "Qwen 3-235B (Local)",
BaseModelShared: BaseModelShared{
- DefaultMaxConvoTokens: 10000, MaxTokens: 131072,
- MaxOutputTokens: 131072, ReservedOutputTokens: 20000,
+ DefaultMaxConvoTokens: 5000, MaxTokens: 40960,
+ MaxOutputTokens: 40960, ReservedOutputTokens: 8192,
PreferredOutputFormat: ModelOutputFormatXml,
},
Providers: []BaseModelUsesProvider{
- {Provider: ModelProviderOllama, ModelName: "ollama_chat/r1-1776:70b"},
+ {Provider: ModelProviderOllama, ModelName: "ollama_chat/qwen3:235b"},
},
},
{
- ModelTag: "perplexity/sonar-reasoning",
- Publisher: ModelPublisherPerplexity,
- Description: "Perplexity Sonar Reasoning",
+ ModelTag: "qwen/qwen3-235b-a22b-cloud",
+ Publisher: ModelPublisherQwen,
+ Description: "Qwen 3-235B A22B (Cloud)",
BaseModelShared: BaseModelShared{
- DefaultMaxConvoTokens: 7500, MaxTokens: 127000,
- MaxOutputTokens: 127000, ReservedOutputTokens: 30000,
+ DefaultMaxConvoTokens: 5000, MaxTokens: 40960,
+ MaxOutputTokens: 40960, ReservedOutputTokens: 8192,
PreferredOutputFormat: ModelOutputFormatXml,
},
- Variants: []BaseModelConfigVariant{
- {VariantTag: "visible", IsDefaultVariant: true, Description: "(reasoning visible)", Overrides: BaseModelShared{IncludeReasoning: true}},
- {VariantTag: "hidden", Description: "(reasoning hidden)", Overrides: BaseModelShared{IncludeReasoning: false}},
- },
Providers: []BaseModelUsesProvider{
- {Provider: ModelProviderPerplexity, ModelName: "sonar-reasoning-online"},
- {Provider: ModelProviderOpenRouter, ModelName: "perplexity/sonar-reasoning"},
+ {Provider: ModelProviderOpenRouter, ModelName: "qwen/qwen3-235b-a22b"},
},
},
{
- ModelTag: "qwen/qwen-2.5-coder-32b-instruct",
+ ModelTag: "qwen/qwen3-32b-local",
Publisher: ModelPublisherQwen,
- Description: "Qwen 2.5 Coder 32B (Instruct)",
+ Description: "Qwen 3-32B (Local)",
BaseModelShared: BaseModelShared{
- DefaultMaxConvoTokens: 10000, MaxTokens: 128000,
- MaxOutputTokens: 8192, ReservedOutputTokens: 8192,
+ DefaultMaxConvoTokens: 5000, MaxTokens: 40960,
+ MaxOutputTokens: 40960, ReservedOutputTokens: 8192,
PreferredOutputFormat: ModelOutputFormatXml,
},
Providers: []BaseModelUsesProvider{
- {Provider: ModelProviderOpenRouter, ModelName: "qwen/qwen-2.5-coder-32b-instruct"},
+ {Provider: ModelProviderOllama, ModelName: "ollama_chat/qwen3:32b"},
},
},
-
{
- ModelTag: "mistral/devstral-small",
- Publisher: ModelPublisherMistral,
- Description: "Mistral Devstral Small",
+ ModelTag: "qwen/qwen3-32b-cloud",
+ Publisher: ModelPublisherQwen,
+ Description: "Qwen 3-32B (Cloud)",
BaseModelShared: BaseModelShared{
- DefaultMaxConvoTokens: 15000, MaxTokens: 128000,
- MaxOutputTokens: 128000, ReservedOutputTokens: 16384,
+ DefaultMaxConvoTokens: 5000, MaxTokens: 40960,
+ MaxOutputTokens: 40960, ReservedOutputTokens: 8192,
PreferredOutputFormat: ModelOutputFormatXml,
},
Providers: []BaseModelUsesProvider{
- {Provider: ModelProviderOllama, ModelName: "ollama_chat/devstral:24b"},
- {Provider: ModelProviderOpenRouter, ModelName: "mistral/devstral-small"},
+ {Provider: ModelProviderOpenRouter, ModelName: "qwen/qwen3-32b"},
},
},
-
{
- ModelTag: "qwen/qwen3-235b",
+ ModelTag: "qwen/qwen3-14b-local",
Publisher: ModelPublisherQwen,
- Description: "Qwen 3-235B",
+ Description: "Qwen 3-14B (Local)",
BaseModelShared: BaseModelShared{
DefaultMaxConvoTokens: 5000, MaxTokens: 40960,
MaxOutputTokens: 40960, ReservedOutputTokens: 8192,
PreferredOutputFormat: ModelOutputFormatXml,
},
Providers: []BaseModelUsesProvider{
- {Provider: ModelProviderOllama, ModelName: "ollama_chat/qwen3:235b"},
- {Provider: ModelProviderOpenRouter, ModelName: "qwen/qwen3-235b-a22b"},
+ {Provider: ModelProviderOllama, ModelName: "ollama_chat/qwen3:14b"},
},
},
{
- ModelTag: "qwen/qwen3-32b",
+ ModelTag: "qwen/qwen3-14b-cloud",
Publisher: ModelPublisherQwen,
- Description: "Qwen 3-32B",
+ Description: "Qwen 3-14B (Cloud)",
BaseModelShared: BaseModelShared{
- DefaultMaxConvoTokens: 15000, MaxTokens: 128000,
- MaxOutputTokens: 16384, ReservedOutputTokens: 16384,
+ DefaultMaxConvoTokens: 5000, MaxTokens: 40960,
+ MaxOutputTokens: 40960, ReservedOutputTokens: 8192,
PreferredOutputFormat: ModelOutputFormatXml,
},
Providers: []BaseModelUsesProvider{
- {Provider: ModelProviderOllama, ModelName: "ollama_chat/qwen3:32b"},
- {Provider: ModelProviderOpenRouter, ModelName: "qwen/qwen3-32b"},
+ {Provider: ModelProviderOpenRouter, ModelName: "qwen/qwen3-14b"},
},
},
{
- ModelTag: "qwen/qwen3-14b",
+ ModelTag: "qwen/qwen3-8b-local",
Publisher: ModelPublisherQwen,
- Description: "Qwen 3-14B",
+ Description: "Qwen 3-8B (Local)",
BaseModelShared: BaseModelShared{
- DefaultMaxConvoTokens: 5000, MaxTokens: 40960,
- MaxOutputTokens: 40960, ReservedOutputTokens: 8192,
+ DefaultMaxConvoTokens: 5000, MaxTokens: 32768,
+ MaxOutputTokens: 32768, ReservedOutputTokens: 8192,
PreferredOutputFormat: ModelOutputFormatXml,
},
Providers: []BaseModelUsesProvider{
- {Provider: ModelProviderOllama, ModelName: "ollama_chat/qwen3:14b"},
- {Provider: ModelProviderOpenRouter, ModelName: "qwen/qwen3-14b"},
+ {Provider: ModelProviderOllama, ModelName: "ollama_chat/qwen3:8b"},
},
},
{
- ModelTag: "qwen/qwen3-8b",
+ ModelTag: "qwen/qwen3-8b-cloud",
Publisher: ModelPublisherQwen,
- Description: "Qwen 3-8B",
+ Description: "Qwen 3-8B (Cloud)",
BaseModelShared: BaseModelShared{
- DefaultMaxConvoTokens: 7500, MaxTokens: 32768,
- MaxOutputTokens: 8192, ReservedOutputTokens: 8192,
+ DefaultMaxConvoTokens: 15000, MaxTokens: 128000,
+ MaxOutputTokens: 20000, ReservedOutputTokens: 20000,
PreferredOutputFormat: ModelOutputFormatXml,
},
Providers: []BaseModelUsesProvider{
- {Provider: ModelProviderOllama, ModelName: "ollama_chat/qwen3:8b"},
{Provider: ModelProviderOpenRouter, ModelName: "qwen/qwen3-8b"},
},
},
+
+ {
+ ModelTag: "mistral/devstral-small",
+ Publisher: ModelPublisherMistral,
+ Description: "Mistral Devstral Small",
+ BaseModelShared: BaseModelShared{
+ DefaultMaxConvoTokens: 15000, MaxTokens: 128000,
+ MaxOutputTokens: 128000, ReservedOutputTokens: 16384,
+ PreferredOutputFormat: ModelOutputFormatXml,
+ },
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderOllama, ModelName: "ollama_chat/devstral:24b"},
+ {Provider: ModelProviderOpenRouter, ModelName: "mistral/devstral-small"},
+ },
+ },
+
+ {
+ ModelTag: "perplexity/r1-1776",
+ Publisher: ModelPublisherPerplexity,
+ Description: "Perplexity R1-1776",
+ BaseModelShared: BaseModelShared{
+ DefaultMaxConvoTokens: 7500, MaxTokens: 128000,
+ MaxOutputTokens: 128000, ReservedOutputTokens: 30000,
+ PreferredOutputFormat: ModelOutputFormatXml,
+ },
+ Variants: []BaseModelConfigVariant{
+ {VariantTag: "visible", IsDefaultVariant: true, Description: "(reasoning visible)", Overrides: BaseModelShared{IncludeReasoning: true}},
+ {VariantTag: "hidden", Description: "(reasoning hidden)", Overrides: BaseModelShared{IncludeReasoning: false}},
+ },
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderPerplexity, ModelName: "r1-1776-online"},
+ {Provider: ModelProviderOpenRouter, ModelName: "perplexity/r1-1776"},
+ },
+ },
+
+ {
+ ModelTag: "perplexity/r1-1776-70b",
+ Publisher: ModelPublisherPerplexity,
+ Description: "Perplexity R1-1776 70B",
+ BaseModelShared: BaseModelShared{
+ DefaultMaxConvoTokens: 10000, MaxTokens: 131072,
+ MaxOutputTokens: 131072, ReservedOutputTokens: 20000,
+ PreferredOutputFormat: ModelOutputFormatXml,
+ },
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderOllama, ModelName: "ollama_chat/r1-1776:70b"},
+ },
+ },
+
+ {
+ ModelTag: "perplexity/sonar-reasoning",
+ Publisher: ModelPublisherPerplexity,
+ Description: "Perplexity Sonar Reasoning",
+ BaseModelShared: BaseModelShared{
+ DefaultMaxConvoTokens: 7500, MaxTokens: 127000,
+ MaxOutputTokens: 127000, ReservedOutputTokens: 30000,
+ PreferredOutputFormat: ModelOutputFormatXml,
+ },
+ Variants: []BaseModelConfigVariant{
+ {VariantTag: "visible", IsDefaultVariant: true, Description: "(reasoning visible)", Overrides: BaseModelShared{IncludeReasoning: true}},
+ {VariantTag: "hidden", Description: "(reasoning hidden)", Overrides: BaseModelShared{IncludeReasoning: false}},
+ },
+ Providers: []BaseModelUsesProvider{
+ {Provider: ModelProviderPerplexity, ModelName: "sonar-reasoning-online"},
+ {Provider: ModelProviderOpenRouter, ModelName: "perplexity/sonar-reasoning"},
+ },
+ },
}
var BuiltInBaseModelsById = map[ModelId]*BaseModelConfigSchema{}
diff --git a/app/shared/ai_models_data_models.go b/app/shared/ai_models_data_models.go
index a5312d67..2fd813a7 100644
--- a/app/shared/ai_models_data_models.go
+++ b/app/shared/ai_models_data_models.go
@@ -41,6 +41,7 @@ type BaseModelShared struct {
ReasoningEffortEnabled bool `json:"reasoningEffortEnabled,omitempty"`
ReasoningEffort ReasoningEffort `json:"reasoningEffort,omitempty"`
IncludeReasoning bool `json:"includeReasoning,omitempty"`
+ HideReasoning bool `json:"hideReasoning,omitempty"`
ReasoningBudget int `json:"reasoningBudget,omitempty"`
SupportsCacheControl bool `json:"supportsCacheControl,omitempty"`
SingleMessageNoSystemPrompt bool `json:"singleMessageNoSystemPrompt,omitempty"`
@@ -98,6 +99,26 @@ type BaseModelConfigVariant struct {
IsDefaultVariant bool `json:"isDefaultVariant"`
}
+func (b *BaseModelConfigSchema) IsLocalOnly() bool {
+ if len(b.Providers) == 0 {
+ return false
+ }
+
+ for _, provider := range b.Providers {
+ builtIn, ok := BuiltInModelProviderConfigs[provider.Provider]
+ if !ok {
+ // has a custom provider—assume not local only
+ return false
+ }
+ if !builtIn.LocalOnly {
+ // has a built-in provider that is not local only
+ return false
+ }
+ }
+
+ return true
+}
+
func (b *BaseModelConfigSchema) ToAvailableModels() []*AvailableModel {
avail := []*AvailableModel{}
for _, provider := range b.Providers {
diff --git a/app/shared/ai_models_providers.go b/app/shared/ai_models_providers.go
index 1769af77..392c1ed3 100644
--- a/app/shared/ai_models_providers.go
+++ b/app/shared/ai_models_providers.go
@@ -56,15 +56,15 @@ var ModelProviderToLiteLLMId = map[ModelProvider]string{
}
var AllModelProviders = []ModelProvider{
- ModelProviderOpenAI,
ModelProviderOpenRouter,
+ ModelProviderOpenAI,
ModelProviderAnthropic,
ModelProviderGoogleAIStudio,
ModelProviderGoogleVertex,
ModelProviderAzureOpenAI,
+ ModelProviderAmazonBedrock,
ModelProviderDeepSeek,
ModelProviderPerplexity,
- ModelProviderAmazonBedrock,
ModelProviderOllama,
ModelProviderCustom,
}