fix(provider): normalize additional provider aliases

crates/tokscale-cli/src/tui/ui/widgets.rs

@@ -207,9 +207,9 @@ fn get_single_provider_display_name(provider: &str) -> String {
         "meta" => "Meta".to_string(),
         "mistral" => "Mistral".to_string(),
         "cohere" => "Cohere".to_string(),
+        "microsoft" => "Microsoft".to_string(),
         "opencode" => "OpenCode".to_string(),
         "owl" => "Owl".to_string(),
-        "github-copilot" => "GitHub Copilot".to_string(),
         "commandcode" => "Command".to_string(),
         "unisound" => "UniSound".to_string(),
         _ => provider.to_string(),
@@ -321,6 +321,9 @@ mod tests {
             ("zai", "Z.AI"),
             ("zai-coding-plan", "Z.AI"),
             ("zai-coding-pln", "Z.AI"),
+            ("z.ai", "Z.AI"),
+            ("bigmodel.cn", "Z.AI"),
+            ("open.bigmodel.cn", "Z.AI"),
             ("zhipuai-coding-plan", "Z.AI"),
             ("zhipu", "Z.AI"),
             ("xiaomi-token-plan-cn", "XiaoMi"),
@@ -329,10 +332,12 @@ mod tests {
             ("minimax-cn-coding-plan", "MiniMax"),
             ("moonshotai", "Kimi"),
             ("moonshot-coding-plan", "Kimi"),
+            ("moonshot-ai", "Kimi"),
             ("kimi", "Kimi"),
             ("kimi-code", "Kimi"),
             ("kimi-for-coding", "Kimi"),
             ("qwen", "Qwen"),
+            ("qwen-portal", "Qwen"),
             ("qwen-coding-plan", "Qwen"),
             ("meituan", "Meituan"),
             ("longcat-coding-plan", "Meituan"),
@@ -343,6 +348,18 @@ mod tests {
             ("alibaba-coding-plan-cn", "Alibaba"),
             ("tencent-coding-plan", "Tencent"),
             ("tecent-coding-plan", "Tencent"),
+            ("xai-oauth", "xAI"),
+            ("grok", "xAI"),
+            ("grok-oauth", "xAI"),
+            ("vertex", "Google"),
+            ("vertex-ai", "Google"),
+            ("google-gemini-cli", "Google"),
+            ("google-antigravity", "Google"),
+            ("azure", "Microsoft"),
+            ("azure-ai", "Microsoft"),
+            ("azure_ai", "Microsoft"),
+            ("github-copilot", "Microsoft"),
+            ("copilot-chat", "Microsoft"),
             ("openai-pro", "OpenAI"),
             ("openai-owl", "Owl"),
             ("openai-owlc", "Owl"),
@@ -384,6 +401,18 @@ mod tests {
             get_provider_display_name("moonshotai, kimi-for-coding"),
             "Kimi"
         );
+        assert_eq!(
+            get_provider_display_name("github-copilot, copilot-chat"),
+            "Microsoft"
+        );
+        assert_eq!(
+            get_provider_display_name("azure, azure-ai, azure_ai, github-copilot"),
+            "Microsoft"
+        );
+        assert_eq!(
+            get_provider_display_name("xai-oauth, grok, grok-oauth"),
+            "xAI"
+        );
         assert_eq!(
             get_provider_display_name("opencode, opencode-go, opencode-zen"),
             "OpenCode"

crates/tokscale-core/src/aggregate/accumulators.rs

@@ -10,7 +10,7 @@ use std::{
 
 use crate::{
     aggregate::keys::{grouped_model_bucket_key, workspace_bucket},
-    normalize_provider_for_grouping, ordered_clients_by_token_contribution, positive_token_total,
+    ordered_clients_by_token_contribution, positive_token_total,
     sessionize::SessionTimeEvent,
     ClientContribution, ClientContributionOrder, DailyContribution, DailyTotals, GraphResult,
     GroupBy, HourlyUsage, ModelPerformance, ModelUsage, MonthlyUsage, SessionContribution,
@@ -45,12 +45,12 @@ impl ModelEntries {
     pub(super) fn push(&mut self, msg: &UnifiedMessage) {
         let group_by = &self.group_by;
         let canonical_model_id = msg.model_id.to_string();
-        let provider = normalize_provider_for_grouping(&msg.provider_id);
+        let provider = msg.provider_id.as_ref();
         let (workspace_group_key, workspace_key, workspace_label) = workspace_bucket(msg);
         let (key, merge_clients) = grouped_model_bucket_key(
             group_by,
             &msg.client,
-            &provider,
+            provider,
             &workspace_group_key,
             &msg.session_id,
             &canonical_model_id,
@@ -82,7 +82,7 @@ impl ModelEntries {
                     None
                 },
                 model: canonical_model_id.clone(),
-                provider: provider.clone(),
+                provider: provider.to_string(),
                 input: 0,
                 output: 0,
                 cache_read: 0,
@@ -362,14 +362,14 @@ impl DailyAcc {
         let client = msg.client.to_string();
         let model = model_id.to_string();
         let key = (client.clone(), model.clone());
-        let provider_id = normalize_provider_for_grouping(&msg.provider_id);
+        let provider_id = msg.provider_id.as_ref();
         let client_entry = self
             .clients
             .entry(key)
             .or_insert_with(|| ClientContribution {
                 client,
                 model_id: model,
-                provider_id: provider_id.clone(),
+                provider_id: provider_id.to_string(),
                 tokens: TokenBreakdown::default(),
                 cost: 0.0,
                 messages: 0,
@@ -552,16 +552,16 @@ impl SessionAcc {
             .saturating_add(msg.tokens.reasoning);
 
         let client = msg.client.to_string();
-        let provider_id = normalize_provider_for_grouping(&msg.provider_id);
+        let provider_id = msg.provider_id.as_ref();
         let model_id = msg.model_id.to_string();
-        let key = (client.clone(), provider_id.clone(), model_id.clone());
+        let key = (client.clone(), provider_id.to_string(), model_id.clone());
         let client_entry = self
             .clients
             .entry(key)
             .or_insert_with(|| ClientContribution {
                 client,
                 model_id,
-                provider_id: provider_id.clone(),
+                provider_id: provider_id.to_string(),
                 tokens: TokenBreakdown::default(),
                 cost: 0.0,
                 messages: 0,

crates/tokscale-core/src/aggregate/engine.rs

@@ -53,7 +53,8 @@ impl AggregationEngine {
 
     /// The per-message fold. `AggregationEngine` consumes finalized local-report
     /// messages. Callers must run `finalize_token_priced_messages` before
-    /// pushing; this layer deliberately does not re-canonicalize model ids.
+    /// pushing; this layer deliberately does not re-canonicalize model or
+    /// provider ids.
     /// Applies the date filter once (mirroring `filter_messages_for_report`)
     /// before dispatching to enabled accumulators.
     pub fn push(&mut self, msg: &UnifiedMessage) {

crates/tokscale-core/src/aggregate/tui.rs

@@ -17,8 +17,8 @@ use crate::{
     aggregate::keys::{
         daily_source_model_key, grouped_model_bucket_key, hourly_model_key, workspace_bucket,
     },
-    normalize_provider_for_grouping, ordered_clients_by_token_contribution, sessions,
-    ClientContributionOrder, GroupBy, ModelPerformance, UnifiedMessage,
+    ordered_clients_by_token_contribution, sessions, ClientContributionOrder, GroupBy,
+    ModelPerformance, UnifiedMessage,
 };
 
 fn positive_unified_token_total(tokens: &crate::TokenBreakdown) -> i64 {
@@ -462,12 +462,12 @@ impl TuiAcc {
     pub(super) fn push(&mut self, msg: &UnifiedMessage) {
         let group_by = &self.group_by;
         let canonical_model_id = msg.model_id.to_string();
-        let provider = normalize_provider_for_grouping(&msg.provider_id);
+        let provider = msg.provider_id.as_ref();
         let (workspace_group_key, workspace_key, workspace_label) = workspace_bucket(msg);
         let (key, merge_clients) = grouped_model_bucket_key(
             group_by,
             &msg.client,
-            &provider,
+            provider,
             &workspace_group_key,
             &msg.session_id,
             &canonical_model_id,
@@ -480,7 +480,7 @@ impl TuiAcc {
             .entry(key.clone())
             .or_insert_with(|| UsageModelEntry {
                 model: canonical_model_id.clone(),
-                provider: provider.clone(),
+                provider: provider.to_string(),
                 client: msg.client.to_string(),
                 workspace_key: if *group_by == GroupBy::WorkspaceModel {
                     workspace_key.clone()
@@ -600,22 +600,22 @@ impl TuiAcc {
                 group_by,
                 &msg.client,
                 &workspace_group_key,
-                &provider,
+                provider,
                 &msg.session_id,
                 &canonical_model_id,
             );
             let model_info = source_entry
                 .models
                 .entry(daily_model_key)
                 .or_insert_with(|| DailyModelInfo {
-                    provider: provider.clone(),
+                    provider: provider.to_string(),
                     display_name: daily_source_model_display_name(
                         group_by,
                         &workspace_label,
                         &msg.session_id,
                         &canonical_model_id,
                     ),
-                    color_key: model_color_key(group_by, &provider, &canonical_model_id),
+                    color_key: model_color_key(group_by, provider, &canonical_model_id),
                     tokens: UsageTokenBreakdown::default(),
                     cost: 0.0,
                     messages: 0,
@@ -647,14 +647,14 @@ impl TuiAcc {
             if msg.is_turn_start {
                 hourly_entry.turn_count += 1;
             }
-            let hkey = hourly_model_key(group_by, &provider, &canonical_model_id);
+            let hkey = hourly_model_key(group_by, provider, &canonical_model_id);
             let hmodel = hourly_entry
                 .models
                 .entry(hkey)
                 .or_insert_with(|| HourlyModelInfo {
-                    provider: provider.clone(),
+                    provider: provider.to_string(),
                     display_name: hourly_model_display_name(group_by, &canonical_model_id),
-                    color_key: model_color_key(group_by, &provider, &canonical_model_id),
+                    color_key: model_color_key(group_by, provider, &canonical_model_id),
                     tokens: UsageTokenBreakdown::default(),
                     cost: 0.0,
                 });
@@ -836,7 +836,7 @@ mod tests {
     }
 
     #[test]
-    fn test_aggregate_messages_model_grouping_normalizes_provider_display_aliases() {
+    fn test_aggregate_messages_model_grouping_uses_finalized_provider_ids() {
         let loader = TuiUsageHarness;
         let usage = loader
             .aggregate_messages(
@@ -853,7 +853,7 @@ mod tests {
                     make_workspace_message(
                         "opencode",
                         "mimo-v2.5-pro",
-                        "xiaomi-token-plan-cn",
+                        "xiaomi",
                         "session-2",
                         2.0,
                         None,
@@ -871,7 +871,7 @@ mod tests {
     }
 
     #[test]
-    fn test_aggregate_messages_client_provider_model_normalizes_provider_display_aliases() {
+    fn test_aggregate_messages_client_provider_model_uses_finalized_provider_ids() {
         let loader = TuiUsageHarness;
         let usage = loader
             .aggregate_messages(
@@ -888,7 +888,7 @@ mod tests {
                     make_workspace_message(
                         "opencode",
                         "mimo-v2.5-pro",
-                        "xiaomi-token-plan-cn",
+                        "xiaomi",
                         "session-2",
                         2.0,
                         None,
@@ -957,7 +957,7 @@ mod tests {
                     make_workspace_message(
                         "opencode",
                         "gpt-5.5",
-                        "azure",
+                        "microsoft",
                         "session-2",
                         2.0,
                         None,
@@ -973,7 +973,7 @@ mod tests {
         let daily_models = &usage.daily[0].source_breakdown["opencode"].models;
         assert_eq!(daily_models.len(), 2);
         assert!(daily_models.contains_key("opencode:openai:gpt-5.5"));
-        assert!(daily_models.contains_key("opencode:azure:gpt-5.5"));
+        assert!(daily_models.contains_key("opencode:microsoft:gpt-5.5"));
         assert!(daily_models
             .values()
             .all(|model| model.display_name == "gpt-5.5"));
@@ -1025,15 +1025,15 @@ mod tests {
     }
 
     #[test]
-    fn test_aggregate_messages_normalizes_moonshot_provider_to_kimi() {
+    fn test_aggregate_messages_uses_finalized_kimi_provider() {
         let loader = TuiUsageHarness;
         let usage = loader
             .aggregate_messages(
                 vec![
                     make_workspace_message(
                         "claude",
                         "kimi-for-coding",
-                        "moonshotai",
+                        "kimi",
                         "session-1",
                         1.0,
                         None,
@@ -1042,7 +1042,7 @@ mod tests {
                     make_workspace_message(
                         "claude",
                         "kimi-for-coding",
-                        "kimi-for-coding",
+                        "kimi",
                         "session-2",
                         2.0,
                         None,
@@ -1444,7 +1444,7 @@ mod tests {
                     UnifiedMessage::new(
                         "claude",
                         "claude-sonnet-4.5",
-                        "github-copilot",
+                        "microsoft",
                         "session-2",
                         1_735_689_600_000,
                         crate::TokenBreakdown {
@@ -1466,7 +1466,7 @@ mod tests {
         assert_eq!(claude.models.len(), 2);
 
         let anthropic_key = "claude:anthropic:claude-sonnet-4.5";
-        let copilot_key = "claude:github-copilot:claude-sonnet-4.5";
+        let copilot_key = "claude:microsoft:claude-sonnet-4.5";
         let anthropic_model = claude.models.get(anthropic_key).unwrap();
         assert_eq!(anthropic_model.display_name, "claude-sonnet-4.5");
         assert_eq!(anthropic_model.provider, "anthropic");
@@ -1475,7 +1475,7 @@ mod tests {
 
         let copilot_model = claude.models.get(copilot_key).unwrap();
         assert_eq!(copilot_model.display_name, "claude-sonnet-4.5");
-        assert_eq!(copilot_model.provider, "github-copilot");
+        assert_eq!(copilot_model.provider, "microsoft");
         assert_eq!(copilot_model.tokens.total(), 30);
         assert_eq!(copilot_model.messages, 1);
     }

crates/tokscale-core/src/lib.rs

@@ -1021,12 +1021,8 @@ fn apply_token_pricing(message: &mut UnifiedMessage, pricing: Option<&pricing::P
 }
 
 fn canonicalize_message_provider(message: &mut UnifiedMessage) {
-    let raw_provider = message.provider_id.trim();
-    let provider = provider_identity::canonical_provider(raw_provider)
-        .or_else(|| {
-            provider_identity::inferred_provider_from_model(&message.model_id).map(str::to_string)
-        })
-        .unwrap_or_else(|| "unknown".to_string());
+    let provider =
+        provider_identity::finalized_provider_id(&message.provider_id, &message.model_id);
     message.provider_id = sessions::intern::intern(&provider);
 }