feat: user cost

go.sum

@@ -55,6 +55,8 @@ github.com/golang/snappy v1.0.0/go.mod h1:/XxbfmMg8lxefKM7IXC3fBNl/7bRcc72aCRzEW
 github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
 github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
 github.com/google/gofuzz v1.0.0/go.mod h1:dBl0BpW6vV/+mYPU4Po3pmUjxk6FQPldtuIdl/M65Eg=
+github.com/google/subcommands v1.2.0 h1:vWQspBTo2nEqTUFita5/KeEWlUL8kQObDFbub/EN9oE=
+github.com/google/subcommands v1.2.0/go.mod h1:ZjhPrFU+Olkh9WazFPsl27BQ4UPiG37m3yTrtFlrHVk=
 github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
 github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
 github.com/google/wire v0.7.0 h1:JxUKI6+CVBgCO2WToKy/nQk0sS+amI9z9EjVmdaocj4=
@@ -160,6 +162,8 @@ golang.org/x/crypto v0.41.0/go.mod h1:pO5AFd7FA68rFak7rOAGVuygIISepHftHnr8dr6+sU
 golang.org/x/exp v0.0.0-20250819193227-8b4c13bb791b h1:DXr+pvt3nC887026GRP39Ej11UATqWDmWuS99x26cD0=
 golang.org/x/exp v0.0.0-20250819193227-8b4c13bb791b/go.mod h1:4QTo5u+SEIbbKW1RacMZq1YEfOBqeXa19JeshGi+zc4=
 golang.org/x/mod v0.6.0-dev.0.20220419223038-86c51ed26bb4/go.mod h1:jJ57K6gSWd91VN4djpZkiMVwK6gcyfeH4XE8wZrZaV4=
+golang.org/x/mod v0.27.0 h1:kb+q2PyFnEADO2IEF935ehFUXlWiNjJWtRNgBLSfbxQ=
+golang.org/x/mod v0.27.0/go.mod h1:rWI627Fq0DEoudcK+MBkNkCe0EetEaDSwJJkCcjpazc=
 golang.org/x/net v0.0.0-20190620200207-3b0461eec859/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
 golang.org/x/net v0.0.0-20210226172049-e18ecbb05110/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
 golang.org/x/net v0.0.0-20220722155237-a158d28d115b/go.mod h1:XRhObCWvk6IyKnWLug+ECip1KBveYUHfp+8e9klMJ9c=
@@ -188,6 +192,8 @@ golang.org/x/text v0.28.0/go.mod h1:U8nCwOR8jO/marOQ0QbDiOngZVEBB7MAiitBuMjXiNU=
 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 v0.36.0 h1:kWS0uv/zsvHEle1LbV5LE8QujrxB3wfQyxHfhOk0Qkg=
+golang.org/x/tools v0.36.0/go.mod h1:WBDiHKJK8YgLHlcQPYQzNCkUxUypCaa5ZegCVutKm+s=
 golang.org/x/xerrors v0.0.0-20190717185122-a985d3407aa7/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
 gonum.org/v1/gonum v0.16.0 h1:5+ul4Swaf3ESvrOnidPp4GZbzf0mxVQpDCYUQE7OJfk=
 gonum.org/v1/gonum v0.16.0/go.mod h1:fef3am4MQ93R2HHpKnLk4/Tbh/s0+wqD5nfa6Pnwy4E=

internal/api/chat/create_conversation_message_stream_v2.go

@@ -281,7 +281,7 @@ func (s *ChatServerV2) CreateConversationMessageStream(
 		APIKey: settings.OpenAIAPIKey,
 	}
 
-	openaiChatHistory, inappChatHistory, err := s.aiClientV2.ChatCompletionStreamV2(ctx, stream, conversation.ID.Hex(), modelSlug, conversation.OpenaiChatHistoryCompletion, llmProvider)
+	openaiChatHistory, inappChatHistory, err := s.aiClientV2.ChatCompletionStreamV2(ctx, stream, conversation.UserID, conversation.ID.Hex(), modelSlug, conversation.OpenaiChatHistoryCompletion, llmProvider)
 	if err != nil {
 		return s.sendStreamError(stream, err)
 	}
@@ -307,7 +307,7 @@ func (s *ChatServerV2) CreateConversationMessageStream(
 			for i, bsonMsg := range conversation.InappChatHistory {
 				protoMessages[i] = mapper.BSONToChatMessageV2(bsonMsg)
 			}
-			title, err := s.aiClientV2.GetConversationTitleV2(ctx, protoMessages, llmProvider)
+			title, err := s.aiClientV2.GetConversationTitleV2(ctx, conversation.UserID, protoMessages, llmProvider)
 			if err != nil {
 				s.logger.Error("Failed to get conversation title", "error", err, "conversationID", conversation.ID.Hex())
 				return

internal/api/grpc.go

@@ -15,6 +15,7 @@ import (
 	chatv2 "paperdebugger/pkg/gen/api/chat/v2"
 	commentv1 "paperdebugger/pkg/gen/api/comment/v1"
 	projectv1 "paperdebugger/pkg/gen/api/project/v1"
+	usagev1 "paperdebugger/pkg/gen/api/usage/v1"
 	userv1 "paperdebugger/pkg/gen/api/user/v1"
 
 	// "github.com/grpc-ecosystem/go-grpc-middleware"
@@ -106,6 +107,7 @@ func NewGrpcServer(
 	userServer userv1.UserServiceServer,
 	projectServer projectv1.ProjectServiceServer,
 	commentServer commentv1.CommentServiceServer,
+	usageServer usagev1.UsageServiceServer,
 ) *GrpcServer {
 	grpcServer := &GrpcServer{}
 	grpcServer.userService = userService
@@ -121,5 +123,6 @@ func NewGrpcServer(
 	userv1.RegisterUserServiceServer(grpcServer.Server, userServer)
 	projectv1.RegisterProjectServiceServer(grpcServer.Server, projectServer)
 	commentv1.RegisterCommentServiceServer(grpcServer.Server, commentServer)
+	usagev1.RegisterUsageServiceServer(grpcServer.Server, usageServer)
 	return grpcServer
 }

internal/api/server.go

@@ -17,6 +17,7 @@ import (
 	commentv1 "paperdebugger/pkg/gen/api/comment/v1"
 	projectv1 "paperdebugger/pkg/gen/api/project/v1"
 	sharedv1 "paperdebugger/pkg/gen/api/shared/v1"
+	usagev1 "paperdebugger/pkg/gen/api/usage/v1"
 	userv1 "paperdebugger/pkg/gen/api/user/v1"
 
 	"github.com/gin-gonic/gin"
@@ -105,6 +106,11 @@ func (s *Server) Run(addr string) {
 		s.logger.Fatalf("failed to register comment service grpc gateway: %v", err)
 		return
 	}
+	err = usagev1.RegisterUsageServiceHandler(context.Background(), mux, client)
+	if err != nil {
+		s.logger.Fatalf("failed to register usage service grpc gateway: %v", err)
+		return
+	}
 
 	s.logger.Infof("[PAPERDEBUGGER] http server listening on %s", addr)
 	s.ginServer.Any("/_pd/api/*path", func(c *gin.Context) { mux.ServeHTTP(c.Writer, c.Request) })

internal/api/usage/get_session_usage.go

@@ -0,0 +1,38 @@
+package usage
+
+import (
+	"context"
+
+	"paperdebugger/internal/libs/contextutil"
+	usagev1 "paperdebugger/pkg/gen/api/usage/v1"
+
+	"google.golang.org/protobuf/types/known/timestamppb"
+)
+
+func (s *UsageServer) GetSessionUsage(
+	ctx context.Context,
+	req *usagev1.GetSessionUsageRequest,
+) (*usagev1.GetSessionUsageResponse, error) {
+	actor, err := contextutil.GetActor(ctx)
+	if err != nil {
+		return nil, err
+	}
+
+	session, err := s.usageService.GetActiveSession(ctx, actor.ID)
+	if err != nil {
+		return nil, err
+	}
+
+	if session == nil {
+		return &usagev1.GetSessionUsageResponse{
+			Session: nil,
+		}, nil
+	}
+
+	return &usagev1.GetSessionUsageResponse{
+		Session: &usagev1.SessionUsage{
+			SessionExpiry: timestamppb.New(session.SessionExpiry.Time()),
+			TotalTokens:   session.TotalTokens,
+		},
+	}, nil
+}

internal/api/usage/get_weekly_usage.go

@@ -0,0 +1,29 @@
+package usage
+
+import (
+	"context"
+
+	"paperdebugger/internal/libs/contextutil"
+	usagev1 "paperdebugger/pkg/gen/api/usage/v1"
+)
+
+func (s *UsageServer) GetWeeklyUsage(
+	ctx context.Context,
+	req *usagev1.GetWeeklyUsageRequest,
+) (*usagev1.GetWeeklyUsageResponse, error) {
+	actor, err := contextutil.GetActor(ctx)
+	if err != nil {
+		return nil, err
+	}
+
+	stats, err := s.usageService.GetWeeklyUsage(ctx, actor.ID)
+	if err != nil {
+		return nil, err
+	}
+
+	return &usagev1.GetWeeklyUsageResponse{
+		Usage: &usagev1.WeeklyUsage{
+			TotalTokens: stats.TotalTokens,
+		},
+	}, nil
+}

internal/api/usage/server.go

@@ -0,0 +1,24 @@
+package usage
+
+import (
+	"paperdebugger/internal/libs/logger"
+	"paperdebugger/internal/services"
+	usagev1 "paperdebugger/pkg/gen/api/usage/v1"
+)
+
+type UsageServer struct {
+	usagev1.UnimplementedUsageServiceServer
+
+	usageService *services.UsageService
+	logger       *logger.Logger
+}
+
+func NewUsageServer(
+	usageService *services.UsageService,
+	logger *logger.Logger,
+) usagev1.UsageServiceServer {
+	return &UsageServer{
+		usageService: usageService,
+		logger:       logger,
+	}
+}

internal/libs/db/db.go

@@ -6,6 +6,7 @@ import (
 
 	"paperdebugger/internal/libs/cfg"
 	"paperdebugger/internal/libs/logger"
+	"paperdebugger/internal/models"
 
 	"go.mongodb.org/mongo-driver/v2/bson"
 	"go.mongodb.org/mongo-driver/v2/mongo"
@@ -43,5 +44,33 @@ func NewDB(cfg *cfg.Cfg, logger *logger.Logger) (*DB, error) {
 	}
 
 	logger.Info("[MONGO] initialized")
-	return &DB{Client: client, cfg: cfg, logger: logger}, nil
+
+	db := &DB{Client: client, cfg: cfg, logger: logger}
+	db.ensureIndexes()
+	return db, nil
+}
+
+// ensureIndexes creates necessary indexes for the database collections.
+func (db *DB) ensureIndexes() {
+	sessions := db.Database("paperdebugger").Collection((models.LLMSession{}).CollectionName())
+
+	// TTL index: auto-delete sessions after 30 days
+	_, err := sessions.Indexes().CreateOne(context.Background(), mongo.IndexModel{
+		Keys:    bson.D{{Key: "session_expiry", Value: 1}},
+		Options: options.Index().SetExpireAfterSeconds(30 * 24 * 60 * 60),
+	})
+	if err != nil {
+		db.logger.Error("Failed to create TTL index on llm_sessions", "error", err)
+	}
+
+	// Compound index for efficient active session lookups
+	_, err = sessions.Indexes().CreateOne(context.Background(), mongo.IndexModel{
+		Keys: bson.D{
+			{Key: "user_id", Value: 1},
+			{Key: "session_expiry", Value: -1},
+		},
+	})
+	if err != nil {
+		db.logger.Error("Failed to create compound index on llm_sessions", "error", err)
+	}
 }

internal/models/usage.go

@@ -0,0 +1,19 @@
+package models
+
+import "go.mongodb.org/mongo-driver/v2/bson"
+
+// LLMSession represents a user's session for tracking LLM usage and token counts.
+type LLMSession struct {
+	ID               bson.ObjectID `bson:"_id"`
+	UserID           bson.ObjectID `bson:"user_id"`
+	SessionStart     bson.DateTime `bson:"session_start"`
+	SessionExpiry    bson.DateTime `bson:"session_expiry"`
+	PromptTokens     int64         `bson:"prompt_tokens"`
+	CompletionTokens int64         `bson:"completion_tokens"`
+	TotalTokens      int64         `bson:"total_tokens"`
+	RequestCount     int64         `bson:"request_count"`
+}
+
+func (s LLMSession) CollectionName() string {
+	return "llm_sessions"
+}

internal/services/toolkit/client/client_v2.go

@@ -20,6 +20,7 @@ type AIClientV2 struct {
 
 	reverseCommentService *services.ReverseCommentService
 	projectService        *services.ProjectService
+	usageService          *services.UsageService
 	cfg                   *cfg.Cfg
 	logger                *logger.Logger
 }
@@ -60,6 +61,7 @@ func NewAIClientV2(
 
 	reverseCommentService *services.ReverseCommentService,
 	projectService *services.ProjectService,
+	usageService *services.UsageService,
 	cfg *cfg.Cfg,
 	logger *logger.Logger,
 ) *AIClientV2 {
@@ -107,6 +109,7 @@ func NewAIClientV2(
 
 		reverseCommentService: reverseCommentService,
 		projectService:        projectService,
+		usageService:          usageService,
 		cfg:                   cfg,
 		logger:                logger,
 	}

internal/services/toolkit/client/completion_v2.go

@@ -4,11 +4,13 @@ import (
 	"context"
 	"encoding/json"
 	"paperdebugger/internal/models"
+	"paperdebugger/internal/services"
 	"paperdebugger/internal/services/toolkit/handler"
 	chatv2 "paperdebugger/pkg/gen/api/chat/v2"
 	"strings"
 
 	"github.com/openai/openai-go/v3"
+	"go.mongodb.org/mongo-driver/v2/bson"
 )
 
 // define []openai.ChatCompletionMessageParamUnion as OpenAIChatHistory
@@ -25,8 +27,8 @@ import (
 //  1. The full chat history sent to the language model (including any tool call results).
 //  2. The incremental chat history visible to the user (including tool call results and assistant responses).
 //  3. An error, if any occurred during the process.
-func (a *AIClientV2) ChatCompletionV2(ctx context.Context, modelSlug string, messages OpenAIChatHistory, llmProvider *models.LLMProviderConfig) (OpenAIChatHistory, AppChatHistory, error) {
-	openaiChatHistory, inappChatHistory, err := a.ChatCompletionStreamV2(ctx, nil, "", modelSlug, messages, llmProvider)
+func (a *AIClientV2) ChatCompletionV2(ctx context.Context, userID bson.ObjectID, modelSlug string, messages OpenAIChatHistory, llmProvider *models.LLMProviderConfig) (OpenAIChatHistory, AppChatHistory, error) {
+	openaiChatHistory, inappChatHistory, err := a.ChatCompletionStreamV2(ctx, nil, userID, "", modelSlug, messages, llmProvider)
 	if err != nil {
 		return nil, nil, err
 	}
@@ -54,7 +56,7 @@ func (a *AIClientV2) ChatCompletionV2(ctx context.Context, modelSlug string, mes
 //   - If tool calls are required, it handles them and appends the results to the chat history, then continues the loop.
 //   - If no tool calls are needed, it appends the assistant's response and exits the loop.
 //   - Finally, it returns the updated chat histories and any error encountered.
-func (a *AIClientV2) ChatCompletionStreamV2(ctx context.Context, callbackStream chatv2.ChatService_CreateConversationMessageStreamServer, conversationId string, modelSlug string, messages OpenAIChatHistory, llmProvider *models.LLMProviderConfig) (OpenAIChatHistory, AppChatHistory, error) {
+func (a *AIClientV2) ChatCompletionStreamV2(ctx context.Context, callbackStream chatv2.ChatService_CreateConversationMessageStreamServer, userID bson.ObjectID, conversationId string, modelSlug string, messages OpenAIChatHistory, llmProvider *models.LLMProviderConfig) (OpenAIChatHistory, AppChatHistory, error) {
 	openaiChatHistory := messages
 	inappChatHistory := AppChatHistory{}
 
@@ -97,7 +99,22 @@ func (a *AIClientV2) ChatCompletionStreamV2(ctx context.Context, callbackStream
 
 			if len(chunk.Choices) == 0 {
 				// Handle usage information
-				// fmt.Printf("Usage: %+v\n", chunk.Usage)
+				if chunk.Usage.TotalTokens > 0 {
+					// Record usage asynchronously to avoid blocking the response
+					go func(usage services.UsageRecord) {
+						bgCtx := context.Background()
+						if err := a.usageService.RecordUsage(bgCtx, usage); err != nil {
+							a.logger.Error("Failed to store usage", "error", err)
+							return
+						}
+
+					}(services.UsageRecord{
+						UserID:           userID,
+						PromptTokens:     chunk.Usage.PromptTokens,
+						CompletionTokens: chunk.Usage.CompletionTokens,
+						TotalTokens:      chunk.Usage.TotalTokens,
+					})
+				}
 				continue
 			}
 
@@ -185,7 +202,6 @@ func (a *AIClientV2) ChatCompletionStreamV2(ctx context.Context, callbackStream
 				// answer_content += chunk.Choices[0].Delta.Content
 				// fmt.Printf("answer_content: %s\n", answer_content)
 				streamHandler.HandleTextDoneItem(chunk, answer_content, reasoning_content)
-				break
 			}
 		}
 

internal/services/toolkit/client/get_citation_keys.go

@@ -241,7 +241,7 @@ func (a *AIClientV2) GetCitationKeys(ctx context.Context, sentence string, userI
 	// Bibliography is placed at the start of the prompt to leverage prompt caching
 	message := fmt.Sprintf("Bibliography: %s\nSentence: %s\nBased on the sentence and bibliography, suggest only the most relevant citation keys separated by commas with no spaces (e.g. key1,key2). Be selective and only include citations that are directly relevant. Avoid suggesting more than 3 citations. If no relevant citations are found, return '%s'.", bibliography, sentence, emptyCitation)
 
-	_, resp, err := a.ChatCompletionV2(ctx, "gpt-5.2", OpenAIChatHistory{
+	_, resp, err := a.ChatCompletionV2(ctx, userId, "gpt-5.2", OpenAIChatHistory{
 		openai.SystemMessage("You are a helpful assistant that suggests relevant citation keys."),
 		openai.UserMessage(message),
 	}, llmProvider)

internal/services/toolkit/client/get_citation_keys_test.go

@@ -25,10 +25,12 @@ func setupTestClient(t *testing.T) (*client.AIClientV2, *services.ProjectService
 	}
 
 	projectService := services.NewProjectService(dbInstance, cfg.GetCfg(), logger.GetLogger())
+	usageService := services.NewUsageService(dbInstance, cfg.GetCfg(), logger.GetLogger())
 	aiClient := client.NewAIClientV2(
 		dbInstance,
 		&services.ReverseCommentService{},
 		projectService,
+		usageService,
 		cfg.GetCfg(),
 		logger.GetLogger(),
 	)

internal/services/toolkit/client/get_conversation_title_v2.go

@@ -11,9 +11,10 @@ import (
 
 	"github.com/openai/openai-go/v3"
 	"github.com/samber/lo"
+	"go.mongodb.org/mongo-driver/v2/bson"
 )
 
-func (a *AIClientV2) GetConversationTitleV2(ctx context.Context, inappChatHistory []*chatv2.Message, llmProvider *models.LLMProviderConfig) (string, error) {
+func (a *AIClientV2) GetConversationTitleV2(ctx context.Context, userID bson.ObjectID, inappChatHistory []*chatv2.Message, llmProvider *models.LLMProviderConfig) (string, error) {
 	messages := lo.Map(inappChatHistory, func(message *chatv2.Message, _ int) string {
 		if _, ok := message.Payload.MessageType.(*chatv2.MessagePayload_Assistant); ok {
 			return fmt.Sprintf("Assistant: %s", message.Payload.GetAssistant().GetContent())
@@ -29,7 +30,7 @@ func (a *AIClientV2) GetConversationTitleV2(ctx context.Context, inappChatHistor
 	message := strings.Join(messages, "\n")
 	message = fmt.Sprintf("%s\nBased on above conversation, generate a short, clear, and descriptive title that summarizes the main topic or purpose of the discussion. The title should be concise, specific, and use natural language. Avoid vague or generic titles. Use abbreviation and short words if possible. Use 3-5 words if possible. Give me the title only, no other text including any other words.", message)
 
-	_, resp, err := a.ChatCompletionV2(ctx, "gpt-5-nano", OpenAIChatHistory{
+	_, resp, err := a.ChatCompletionV2(ctx, userID, "gpt-5-nano", OpenAIChatHistory{
 		openai.SystemMessage("You are a helpful assistant that generates a title for a conversation."),
 		openai.UserMessage(message),
 	}, llmProvider)

internal/services/toolkit/client/utils_v2.go

@@ -74,6 +74,9 @@ func getDefaultParamsV2(modelSlug string, toolRegistry *registry.ToolRegistryV2)
 				Tools:               toolRegistry.GetTools(),
 				ParallelToolCalls:   openaiv3.Bool(true),
 				Store:               openaiv3.Bool(false),
+				StreamOptions: openaiv3.ChatCompletionStreamOptionsParam{
+					IncludeUsage: openaiv3.Bool(true),
+				},
 			}
 		}
 	}
@@ -85,6 +88,9 @@ func getDefaultParamsV2(modelSlug string, toolRegistry *registry.ToolRegistryV2)
 		Tools:               toolRegistry.GetTools(), // Tool registration is managed centrally by the registry
 		ParallelToolCalls:   openaiv3.Bool(true),
 		Store:               openaiv3.Bool(false), // Must set to false, because we are construct our own chat history.
+		StreamOptions: openaiv3.ChatCompletionStreamOptionsParam{
+			IncludeUsage: openaiv3.Bool(true),
+		},
 	}
 }