refactor(models): use models.dev API for model discovery

Author: itzlambda

README.md

@@ -66,7 +66,7 @@ rullm --system "You are a helpful assistant." "Summarize this text"
 # List available models (shows only chat models, with your aliases)
 rullm models list
 
-# Update model list for all providers with API keys
+# Update model list from models.dev (no API keys required)
 rullm models update
 
 # Manage aliases
@@ -129,4 +129,4 @@ source <(COMPLETE=bash ./target/debug/rullm)
 
 # zsh
 source <(COMPLETE=zsh ./target/debug/rullm)
-```
+```

crates/rullm-cli/src/cli_client.rs

@@ -486,18 +486,6 @@ impl CliClient {
         }
     }
 
-    /// Get available models for the provider
-    pub async fn available_models(&self) -> Result<Vec<String>, LlmError> {
-        match self {
-            Self::OpenAI { client, .. } => client.list_models().await,
-            Self::Anthropic { client, .. } => client.list_models().await,
-            Self::Google { client, .. } => client.list_models().await,
-            Self::Groq { client, .. } | Self::OpenRouter { client, .. } => {
-                client.available_models().await
-            }
-        }
-    }
-
     /// Get provider name
     pub fn provider_name(&self) -> &'static str {
         match self {

crates/rullm-cli/src/commands/mod.rs

@@ -37,7 +37,7 @@ const CHAT_EXAMPLES: &str = r#"EXAMPLES:
 
 const MODELS_EXAMPLES: &str = r#"EXAMPLES:
   rullm models list                        # List cached models
-  rullm models update -m openai/gpt-4      # Fetch OpenAI models
+  rullm models update                      # Fetch latest models
   rullm models default openai/gpt-4o       # Set default model
   rullm models clear                       # Clear model cache"#;
 

crates/rullm-cli/src/commands/models.rs

@@ -1,14 +1,13 @@
-use crate::cli_client::CliClient;
 use anyhow::Result;
 use chrono::Utc;
 use clap::{Args, Subcommand};
-use rullm_core::LlmError;
+use serde::Deserialize;
+use std::collections::HashMap;
 use strum::IntoEnumIterator;
 
 use crate::{
     aliases::UserAliasConfig,
     args::{Cli, CliConfig},
-    client,
     commands::{ModelsCache, format_duration},
     constants::{ALIASES_CONFIG_FILE, MODEL_FILE_NAME},
     output::OutputLevel,
@@ -23,14 +22,14 @@ pub struct ModelsArgs {
 
 #[derive(Subcommand)]
 pub enum ModelsAction {
-    /// List available models for the current provider (default)
+    /// List cached models
     List,
     /// Set a default model that will be used when --model is not supplied
     Default {
         /// Model identifier in the form provider:model-name (e.g. openai:gpt-4o)
         model: Option<String>,
     },
-    /// Fetch fresh models from all providers with available API keys and update local cache
+    /// Fetch fresh models from models.dev and update local cache
     Update,
     /// Clear the local models cache
     Clear,
@@ -41,7 +40,7 @@ impl ModelsArgs {
         &self,
         output_level: OutputLevel,
         cli_config: &mut CliConfig,
-        cli: &Cli,
+        _cli: &Cli,
     ) -> Result<()> {
         match &self.action {
             ModelsAction::List => {
@@ -67,34 +66,28 @@ impl ModelsArgs {
                 }
             }
             ModelsAction::Update => {
-                // List of supported providers
-                let providers = Provider::iter();
-                let mut updated = vec![];
-                let mut skipped = vec![];
-
-                for provider in providers {
-                    let provider = format!("{provider}");
-                    // Try to create a client for this provider
-                    let model_hint = format!("{provider}:dummy"); // dummy model name, just to get the client
-                    let client = match client::from_model(&model_hint, cli, cli_config).await {
-                        Ok(c) => c,
-                        Err(_) => {
-                            skipped.push(provider);
-                            continue;
-                        }
-                    };
-                    match update_models(cli_config, &client, output_level).await {
-                        Ok(_) => updated.push(provider),
-                        Err(_) => skipped.push(provider),
-                    }
+                let supported: Vec<&str> = Provider::iter().map(|p| p.models_dev_id()).collect();
+
+                crate::output::progress("Fetching models from models.dev...", output_level);
+
+                let models = fetch_models_from_models_dev(&supported).await?;
+                if models.is_empty() {
+                    anyhow::bail!("No models returned by models.dev");
                 }
 
-                if !skipped.is_empty() {
-                    crate::output::note(
-                        &format!("Skipped (no API key or error): {}", skipped.join(", ")),
-                        output_level,
-                    );
+                let cache = ModelsCache::new(models);
+                let path = cli_config.data_base_path.join(MODEL_FILE_NAME);
+
+                if let Some(parent) = path.parent() {
+                    std::fs::create_dir_all(parent)?;
                 }
+
+                std::fs::write(&path, serde_json::to_string_pretty(&cache)?)?;
+
+                crate::output::success(
+                    &format!("Updated {} models", cache.models.len()),
+                    output_level,
+                );
             }
             ModelsAction::Clear => {
                 clear_models_cache(cli_config, output_level)?;
@@ -216,85 +209,6 @@ pub fn clear_models_cache(cli_config: &CliConfig, output_level: OutputLevel) ->
     Ok(())
 }
 
-pub async fn update_models(
-    cli_config: &mut CliConfig,
-    client: &CliClient,
-    output_level: OutputLevel,
-) -> Result<(), LlmError> {
-    crate::output::progress(
-        &format!(
-            "Fetching models from {}...",
-            crate::output::format_provider(client.provider_name())
-        ),
-        output_level,
-    );
-
-    let mut models = client.available_models().await.map_err(|e| {
-        crate::output::error(&format!("Failed to fetch models: {e}"), output_level);
-        e
-    })?;
-
-    if models.is_empty() {
-        crate::output::error("No models returned by provider", output_level);
-        return Err(LlmError::model(
-            "No models returned by provider".to_string(),
-        ));
-    }
-
-    models.sort();
-    models.dedup();
-
-    _cache_models(cli_config, client.provider_name(), &models).map_err(|e| {
-        crate::output::error(&format!("Failed to update models cache: {e}"), output_level);
-        LlmError::unknown(e.to_string())
-    })?;
-
-    crate::output::success(
-        &format!(
-            "Updated {} models for {}",
-            models.len(),
-            client.provider_name()
-        ),
-        output_level,
-    );
-
-    Ok(())
-}
-
-fn _cache_models(cli_config: &CliConfig, provider_name: &str, models: &[String]) -> Result<()> {
-    use std::fs;
-
-    let path = cli_config.data_base_path.join(MODEL_FILE_NAME);
-    // TODO: we shouldn't need to do this here, this should be done while cli_config is created
-    // TODO: Remove if we already do this.
-    if let Some(parent) = path.parent() {
-        fs::create_dir_all(parent)?;
-    }
-
-    // Load existing cache if present
-    let mut entries = if let Ok(Some(cache)) = load_models_cache(cli_config) {
-        cache.models
-    } else {
-        Vec::new()
-    };
-
-    // Remove all entries for this provider
-    let prefix = format!("{}:", provider_name.to_lowercase());
-    entries.retain(|m| !m.starts_with(&prefix));
-
-    // Add new models for this provider
-    let new_entries: Vec<String> = models
-        .iter()
-        .map(|m| format!("{}:{}", provider_name.to_lowercase(), m))
-        .collect();
-    entries.extend(new_entries);
-
-    let cache = ModelsCache::new(entries);
-    let json = serde_json::to_string_pretty(&cache)?;
-    fs::write(path, json)?;
-    Ok(())
-}
-
 pub(crate) fn load_models_cache(cli_config: &CliConfig) -> Result<Option<ModelsCache>> {
     use std::fs;
 
@@ -314,3 +228,35 @@ pub(crate) fn load_models_cache(cli_config: &CliConfig) -> Result<Option<ModelsC
     // Old format doesn't have timestamp info
     Ok(None)
 }
+
+#[derive(Deserialize)]
+struct ModelsDevProvider {
+    models: HashMap<String, ModelsDevModel>,
+}
+
+#[derive(Deserialize)]
+struct ModelsDevModel {
+    #[serde(default)]
+    id: Option<String>,
+}
+
+async fn fetch_models_from_models_dev(supported_providers: &[&str]) -> Result<Vec<String>> {
+    let response = reqwest::get("https://models.dev/api.json")
+        .await?
+        .error_for_status()?;
+    let providers: HashMap<String, ModelsDevProvider> = response.json().await?;
+
+    let mut all_models = Vec::new();
+    for provider_id in supported_providers {
+        if let Some(provider) = providers.get(*provider_id) {
+            for (model_id, model) in &provider.models {
+                let id = model.id.as_deref().unwrap_or(model_id);
+                all_models.push(format!("{provider_id}:{id}"));
+            }
+        }
+    }
+
+    all_models.sort();
+    all_models.dedup();
+    Ok(all_models)
+}

crates/rullm-cli/src/provider.rs

@@ -90,4 +90,14 @@ impl Provider {
             Provider::Google => "GOOGLE_AI_API_KEY",
         }
     }
+
+    pub fn models_dev_id(&self) -> &'static str {
+        match self {
+            Provider::OpenAI => "openai",
+            Provider::Groq => "groq",
+            Provider::OpenRouter => "openrouter",
+            Provider::Anthropic => "anthropic",
+            Provider::Google => "google",
+        }
+    }
 }

crates/rullm-core/examples/README.md

@@ -348,7 +348,6 @@ match provider.chat_completion(request).await {
 
 - **`provider.chat_completion(request)`** - Send chat completion
 - **`provider.health_check()`** - Test API connectivity
-- **`provider.available_models()`** - Get supported models
 - **`config.validate()`** - Validate configuration
 
 ### Supported Models
@@ -391,7 +390,7 @@ match provider.chat_completion(request).await {
 
 ## Test All Providers (`test_all_providers.rs`)
 
-Comprehensive test that validates all LLM providers and their `available_models` functionality:
+Comprehensive test that validates all LLM providers with health checks:
 
 ```bash
 # Set up your API keys
@@ -405,31 +404,26 @@ cargo run --example test_all_providers
 
 **Features:**
 - Tests OpenAI, Anthropic, and Google providers
-- Calls `available_models()` for each provider
-- Validates expected model patterns
 - Performs health checks
 - Provides detailed success/failure reporting
 - Gracefully handles missing API keys
 
 **Sample Output:**
 ```
-🚀 Testing All LLM Providers and Their Available Models
+🚀 Testing All LLM Providers
 
 🔍 Testing OpenAI Provider...
-   Provider name: openai
    Health check: ✅ Passed
-   Expected model 'gpt-4': ✅ Found
-   Expected model 'gpt-3.5-turbo': ✅ Found
-✅ OpenAI: Found 5 models
+✅ OpenAI: Health check passed
 
 📊 SUMMARY:
-┌─────────────┬────────┬─────────────┐
-│ Provider    │ Status │ Models      │
-├─────────────┼────────┼─────────────┤
-│ OpenAI      │ ✅ Pass │ 5 models    │
-│ Anthropic   │ ✅ Pass │ 5 models    │
-│ Google      │ ✅ Pass │ 5 models    │
-└─────────────┴────────┴─────────────┘
+┌─────────────┬────────┐
+│ Provider    │ Status │
+├─────────────┼────────┤
+│ OpenAI      │ ✅ Pass │
+│ Anthropic   │ ✅ Pass │
+│ Google      │ ✅ Pass │
+└─────────────┴────────┘
 
 🎉 All providers are working correctly!
 ```
@@ -438,4 +432,4 @@ Use this example for:
 - Verifying your API keys work
 - Testing network connectivity
 - Validating provider implementations
-- CI/CD pipeline health checks 
+- CI/CD pipeline health checks 

crates/rullm-core/examples/google_simple.rs

@@ -125,17 +125,7 @@ async fn main() -> Result<(), Box<dyn std::error::Error>> {
         }
     }
 
-    // 7. List models
-    println!("\n📋 Available models:");
-    let models = client.list_models().await?;
-    for (i, model) in models.iter().take(5).enumerate() {
-        println!("  {}. {}", i + 1, model);
-    }
-    if models.len() > 5 {
-        println!("  ... and {} more", models.len() - 5);
-    }
-
-    // 8. Health check
+    // 7. Health check
     match client.health_check().await {
         Ok(_) => println!("\n✅ Google AI is healthy"),
         Err(e) => println!("\n❌ Health check failed: {e}"),

crates/rullm-core/examples/openai_config.rs

@@ -72,20 +72,6 @@ async fn main() -> Result<(), Box<dyn std::error::Error>> {
                                 Err(e) => println!("   ❌ Health check failed: {e}"),
                             }
 
-                            // Get available models
-                            match client.list_models().await {
-                                Ok(models) => {
-                                    println!("   Available models (first 5):");
-                                    for (i, model) in models.iter().take(5).enumerate() {
-                                        println!("     {}. {}", i + 1, model);
-                                    }
-                                    if models.len() > 5 {
-                                        println!("     ... and {} more", models.len() - 5);
-                                    }
-                                }
-                                Err(e) => println!("   ❌ Error getting models: {e}"),
-                            }
-
                             // Make a simple request
                             println!("\n   Testing chat completion...");
                             let mut test_request = ChatCompletionRequest::new(