rework llm

4 files changed, 1078 insertions, 0 deletions

diff --git a/internal/llm/provider/anthropic.go b/internal/llm/provider/anthropic.go
new file mode 100644
index 000000000..63a68b92b
--- /dev/null
+++ b/internal/llm/provider/anthropic.go
@@ -0,0 +1,309 @@
+package provider
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"strings"
+
+	"github.com/anthropics/anthropic-sdk-go"
+	"github.com/anthropics/anthropic-sdk-go/option"
+	"github.com/kujtimiihoxha/termai/internal/llm/models"
+	"github.com/kujtimiihoxha/termai/internal/llm/tools"
+	"github.com/kujtimiihoxha/termai/internal/message"
+)
+
+type anthropicProvider struct {
+	client        anthropic.Client
+	model         models.Model
+	maxTokens     int64
+	apiKey        string
+	systemMessage string
+}
+
+type AnthropicOption func(*anthropicProvider)
+
+func WithAnthropicSystemMessage(message string) AnthropicOption {
+	return func(a *anthropicProvider) {
+		a.systemMessage = message
+	}
+}
+
+func WithAnthropicMaxTokens(maxTokens int64) AnthropicOption {
+	return func(a *anthropicProvider) {
+		a.maxTokens = maxTokens
+	}
+}
+
+func WithAnthropicModel(model models.Model) AnthropicOption {
+	return func(a *anthropicProvider) {
+		a.model = model
+	}
+}
+
+func WithAnthropicKey(apiKey string) AnthropicOption {
+	return func(a *anthropicProvider) {
+		a.apiKey = apiKey
+	}
+}
+
+func NewAnthropicProvider(opts ...AnthropicOption) (Provider, error) {
+	provider := &anthropicProvider{
+		maxTokens: 1024,
+	}
+
+	for _, opt := range opts {
+		opt(provider)
+	}
+
+	if provider.systemMessage == "" {
+		return nil, errors.New("system message is required")
+	}
+
+	provider.client = anthropic.NewClient(option.WithAPIKey(provider.apiKey))
+	return provider, nil
+}
+
+func (a *anthropicProvider) SendMessages(ctx context.Context, messages []message.Message, tools []tools.BaseTool) (*ProviderResponse, error) {
+	anthropicMessages := a.convertToAnthropicMessages(messages)
+	anthropicTools := a.convertToAnthropicTools(tools)
+
+	response, err := a.client.Messages.New(ctx, anthropic.MessageNewParams{
+		Model:       anthropic.Model(a.model.APIModel),
+		MaxTokens:   a.maxTokens,
+		Temperature: anthropic.Float(0),
+		Messages:    anthropicMessages,
+		Tools:       anthropicTools,
+		System: []anthropic.TextBlockParam{
+			{
+				Text: a.systemMessage,
+				CacheControl: anthropic.CacheControlEphemeralParam{
+					Type: "ephemeral",
+				},
+			},
+		},
+	})
+	if err != nil {
+		return nil, err
+	}
+
+	content := ""
+	for _, block := range response.Content {
+		if text, ok := block.AsAny().(anthropic.TextBlock); ok {
+			content += text.Text
+		}
+	}
+
+	toolCalls := a.extractToolCalls(response.Content)
+	tokenUsage := a.extractTokenUsage(response.Usage)
+
+	return &ProviderResponse{
+		Content:   content,
+		ToolCalls: toolCalls,
+		Usage:     tokenUsage,
+	}, nil
+}
+
+func (a *anthropicProvider) StreamResponse(ctx context.Context, messages []message.Message, tools []tools.BaseTool) (<-chan ProviderEvent, error) {
+	anthropicMessages := a.convertToAnthropicMessages(messages)
+	anthropicTools := a.convertToAnthropicTools(tools)
+
+	var thinkingParam anthropic.ThinkingConfigParamUnion
+	lastMessage := messages[len(messages)-1]
+	temperature := anthropic.Float(0)
+	if lastMessage.Role == message.User && strings.Contains(strings.ToLower(lastMessage.Content), "think") {
+		thinkingParam = anthropic.ThinkingConfigParamUnion{
+			OfThinkingConfigEnabled: &anthropic.ThinkingConfigEnabledParam{
+				BudgetTokens: int64(float64(a.maxTokens) * 0.8),
+				Type:         "enabled",
+			},
+		}
+		temperature = anthropic.Float(1)
+	}
+
+	stream := a.client.Messages.NewStreaming(ctx, anthropic.MessageNewParams{
+		Model:       anthropic.Model(a.model.APIModel),
+		MaxTokens:   a.maxTokens,
+		Temperature: temperature,
+		Messages:    anthropicMessages,
+		Tools:       anthropicTools,
+		Thinking:    thinkingParam,
+		System: []anthropic.TextBlockParam{
+			{
+				Text: a.systemMessage,
+				CacheControl: anthropic.CacheControlEphemeralParam{
+					Type: "ephemeral",
+				},
+			},
+		},
+	})
+
+	eventChan := make(chan ProviderEvent)
+
+	go func() {
+		defer close(eventChan)
+
+		accumulatedMessage := anthropic.Message{}
+
+		for stream.Next() {
+			event := stream.Current()
+			err := accumulatedMessage.Accumulate(event)
+			if err != nil {
+				eventChan <- ProviderEvent{Type: EventError, Error: err}
+				return
+			}
+
+			switch event := event.AsAny().(type) {
+			case anthropic.ContentBlockStartEvent:
+				eventChan <- ProviderEvent{Type: EventContentStart}
+
+			case anthropic.ContentBlockDeltaEvent:
+				if event.Delta.Type == "thinking_delta" && event.Delta.Thinking != "" {
+					eventChan <- ProviderEvent{
+						Type:     EventThinkingDelta,
+						Thinking: event.Delta.Thinking,
+					}
+				} else if event.Delta.Type == "text_delta" && event.Delta.Text != "" {
+					eventChan <- ProviderEvent{
+						Type:    EventContentDelta,
+						Content: event.Delta.Text,
+					}
+				}
+
+			case anthropic.ContentBlockStopEvent:
+				eventChan <- ProviderEvent{Type: EventContentStop}
+
+			case anthropic.MessageStopEvent:
+				content := ""
+				for _, block := range accumulatedMessage.Content {
+					if text, ok := block.AsAny().(anthropic.TextBlock); ok {
+						content += text.Text
+					}
+				}
+
+				toolCalls := a.extractToolCalls(accumulatedMessage.Content)
+				tokenUsage := a.extractTokenUsage(accumulatedMessage.Usage)
+
+				eventChan <- ProviderEvent{
+					Type: EventComplete,
+					Response: &ProviderResponse{
+						Content:   content,
+						ToolCalls: toolCalls,
+						Usage:     tokenUsage,
+					},
+				}
+			}
+		}
+
+		if stream.Err() != nil {
+			eventChan <- ProviderEvent{Type: EventError, Error: stream.Err()}
+		}
+	}()
+
+	return eventChan, nil
+}
+
+func (a *anthropicProvider) extractToolCalls(content []anthropic.ContentBlockUnion) []message.ToolCall {
+	var toolCalls []message.ToolCall
+
+	for _, block := range content {
+		switch variant := block.AsAny().(type) {
+		case anthropic.ToolUseBlock:
+			toolCall := message.ToolCall{
+				ID:    variant.ID,
+				Name:  variant.Name,
+				Input: string(variant.Input),
+				Type:  string(variant.Type),
+			}
+			toolCalls = append(toolCalls, toolCall)
+		}
+	}
+
+	return toolCalls
+}
+
+func (a *anthropicProvider) extractTokenUsage(usage anthropic.Usage) TokenUsage {
+	return TokenUsage{
+		InputTokens:         usage.InputTokens,
+		OutputTokens:        usage.OutputTokens,
+		CacheCreationTokens: usage.CacheCreationInputTokens,
+		CacheReadTokens:     usage.CacheReadInputTokens,
+	}
+}
+
+func (a *anthropicProvider) convertToAnthropicTools(tools []tools.BaseTool) []anthropic.ToolUnionParam {
+	anthropicTools := make([]anthropic.ToolUnionParam, len(tools))
+
+	for i, tool := range tools {
+		info := tool.Info()
+		toolParam := anthropic.ToolParam{
+			Name:        info.Name,
+			Description: anthropic.String(info.Description),
+			InputSchema: anthropic.ToolInputSchemaParam{
+				Properties: info.Parameters,
+			},
+		}
+
+		if i == len(tools)-1 {
+			toolParam.CacheControl = anthropic.CacheControlEphemeralParam{
+				Type: "ephemeral",
+			}
+		}
+
+		anthropicTools[i] = anthropic.ToolUnionParam{OfTool: &toolParam}
+	}
+
+	return anthropicTools
+}
+
+func (a *anthropicProvider) convertToAnthropicMessages(messages []message.Message) []anthropic.MessageParam {
+	anthropicMessages := make([]anthropic.MessageParam, len(messages))
+	cachedBlocks := 0
+
+	for i, msg := range messages {
+		switch msg.Role {
+		case message.User:
+			content := anthropic.NewTextBlock(msg.Content)
+			if cachedBlocks < 2 {
+				content.OfRequestTextBlock.CacheControl = anthropic.CacheControlEphemeralParam{
+					Type: "ephemeral",
+				}
+				cachedBlocks++
+			}
+			anthropicMessages[i] = anthropic.NewUserMessage(content)
+
+		case message.Assistant:
+			blocks := []anthropic.ContentBlockParamUnion{}
+			if msg.Content != "" {
+				content := anthropic.NewTextBlock(msg.Content)
+				if cachedBlocks < 2 {
+					content.OfRequestTextBlock.CacheControl = anthropic.CacheControlEphemeralParam{
+						Type: "ephemeral",
+					}
+					cachedBlocks++
+				}
+				blocks = append(blocks, content)
+			}
+
+			for _, toolCall := range msg.ToolCalls {
+				var inputMap map[string]any
+				err := json.Unmarshal([]byte(toolCall.Input), &inputMap)
+				if err != nil {
+					continue
+				}
+				blocks = append(blocks, anthropic.ContentBlockParamOfRequestToolUseBlock(toolCall.ID, inputMap, toolCall.Name))
+			}
+
+			anthropicMessages[i] = anthropic.NewAssistantMessage(blocks...)
+
+		case message.Tool:
+			results := make([]anthropic.ContentBlockParamUnion, len(msg.ToolResults))
+			for i, toolResult := range msg.ToolResults {
+				results[i] = anthropic.NewToolResultBlock(toolResult.ToolCallID, toolResult.Content, toolResult.IsError)
+			}
+			anthropicMessages[i] = anthropic.NewUserMessage(results...)
+		}
+	}
+
+	return anthropicMessages
+}
diff --git a/internal/llm/provider/gemini.go b/internal/llm/provider/gemini.go
new file mode 100644
index 000000000..6b252b581
--- /dev/null
+++ b/internal/llm/provider/gemini.go
@@ -0,0 +1,443 @@
+package provider
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"log"
+
+	"github.com/google/generative-ai-go/genai"
+	"github.com/google/uuid"
+	"github.com/kujtimiihoxha/termai/internal/llm/models"
+	"github.com/kujtimiihoxha/termai/internal/llm/tools"
+	"github.com/kujtimiihoxha/termai/internal/message"
+	"google.golang.org/api/googleapi"
+	"google.golang.org/api/iterator"
+	"google.golang.org/api/option"
+)
+
+type geminiProvider struct {
+	client        *genai.Client
+	model         models.Model
+	maxTokens     int32
+	apiKey        string
+	systemMessage string
+}
+
+type GeminiOption func(*geminiProvider)
+
+func NewGeminiProvider(ctx context.Context, opts ...GeminiOption) (Provider, error) {
+	provider := &geminiProvider{
+		maxTokens: 5000,
+	}
+
+	for _, opt := range opts {
+		opt(provider)
+	}
+
+	if provider.systemMessage == "" {
+		return nil, errors.New("system message is required")
+	}
+
+	client, err := genai.NewClient(ctx, option.WithAPIKey(provider.apiKey))
+	if err != nil {
+		return nil, err
+	}
+	provider.client = client
+
+	return provider, nil
+}
+
+func WithGeminiSystemMessage(message string) GeminiOption {
+	return func(p *geminiProvider) {
+		p.systemMessage = message
+	}
+}
+
+func WithGeminiMaxTokens(maxTokens int32) GeminiOption {
+	return func(p *geminiProvider) {
+		p.maxTokens = maxTokens
+	}
+}
+
+func WithGeminiModel(model models.Model) GeminiOption {
+	return func(p *geminiProvider) {
+		p.model = model
+	}
+}
+
+func WithGeminiKey(apiKey string) GeminiOption {
+	return func(p *geminiProvider) {
+		p.apiKey = apiKey
+	}
+}
+
+func (p *geminiProvider) Close() {
+	if p.client != nil {
+		p.client.Close()
+	}
+}
+
+// convertToGeminiHistory converts the message history to Gemini's format
+func (p *geminiProvider) convertToGeminiHistory(messages []message.Message) []*genai.Content {
+	var history []*genai.Content
+
+	for _, msg := range messages {
+		switch msg.Role {
+		case message.User:
+			history = append(history, &genai.Content{
+				Parts: []genai.Part{genai.Text(msg.Content)},
+				Role:  "user",
+			})
+		case message.Assistant:
+			content := &genai.Content{
+				Role:  "model",
+				Parts: []genai.Part{},
+			}
+
+			// Handle regular content
+			if msg.Content != "" {
+				content.Parts = append(content.Parts, genai.Text(msg.Content))
+			}
+
+			// Handle tool calls if any
+			if len(msg.ToolCalls) > 0 {
+				for _, call := range msg.ToolCalls {
+					args, _ := parseJsonToMap(call.Input)
+					content.Parts = append(content.Parts, genai.FunctionCall{
+						Name: call.Name,
+						Args: args,
+					})
+				}
+			}
+
+			history = append(history, content)
+		case message.Tool:
+			for _, result := range msg.ToolResults {
+				// Parse response content to map if possible
+				response := map[string]interface{}{"result": result.Content}
+				parsed, err := parseJsonToMap(result.Content)
+				if err == nil {
+					response = parsed
+				}
+				var toolCall message.ToolCall
+				for _, msg := range messages {
+					if msg.Role == message.Assistant {
+						for _, call := range msg.ToolCalls {
+							if call.ID == result.ToolCallID {
+								toolCall = call
+								break
+							}
+						}
+					}
+				}
+
+				history = append(history, &genai.Content{
+					Parts: []genai.Part{genai.FunctionResponse{
+						Name:     toolCall.Name,
+						Response: response,
+					}},
+					Role: "function",
+				})
+			}
+		}
+	}
+
+	return history
+}
+
+// convertToolsToGeminiFunctionDeclarations converts tool definitions to Gemini's function declarations
+func (p *geminiProvider) convertToolsToGeminiFunctionDeclarations(tools []tools.BaseTool) []*genai.FunctionDeclaration {
+	declarations := make([]*genai.FunctionDeclaration, len(tools))
+
+	for i, tool := range tools {
+		info := tool.Info()
+
+		// Convert parameters to genai.Schema format
+		properties := make(map[string]*genai.Schema)
+		for name, param := range info.Parameters {
+			// Try to extract type and description from the parameter
+			paramMap, ok := param.(map[string]interface{})
+			if !ok {
+				// Default to string if unable to determine type
+				properties[name] = &genai.Schema{Type: genai.TypeString}
+				continue
+			}
+
+			schemaType := genai.TypeString // Default
+			var description string
+			var itemsTypeSchema *genai.Schema
+			if typeVal, found := paramMap["type"]; found {
+				if typeStr, ok := typeVal.(string); ok {
+					switch typeStr {
+					case "string":
+						schemaType = genai.TypeString
+					case "number":
+						schemaType = genai.TypeNumber
+					case "integer":
+						schemaType = genai.TypeInteger
+					case "boolean":
+						schemaType = genai.TypeBoolean
+					case "array":
+						schemaType = genai.TypeArray
+						items, found := paramMap["items"]
+						if found {
+							itemsMap, ok := items.(map[string]interface{})
+							if ok {
+								itemsType, found := itemsMap["type"]
+								if found {
+									itemsTypeStr, ok := itemsType.(string)
+									if ok {
+										switch itemsTypeStr {
+										case "string":
+											itemsTypeSchema = &genai.Schema{
+												Type: genai.TypeString,
+											}
+										case "number":
+											itemsTypeSchema = &genai.Schema{
+												Type: genai.TypeNumber,
+											}
+										case "integer":
+											itemsTypeSchema = &genai.Schema{
+												Type: genai.TypeInteger,
+											}
+										case "boolean":
+											itemsTypeSchema = &genai.Schema{
+												Type: genai.TypeBoolean,
+											}
+										}
+									}
+								}
+							}
+						}
+					case "object":
+						schemaType = genai.TypeObject
+						if _, found := paramMap["properties"]; !found {
+							continue
+						}
+						// TODO: Add support for other types
+					}
+				}
+			}
+
+			if desc, found := paramMap["description"]; found {
+				if descStr, ok := desc.(string); ok {
+					description = descStr
+				}
+			}
+
+			properties[name] = &genai.Schema{
+				Type:        schemaType,
+				Description: description,
+				Items:       itemsTypeSchema,
+			}
+		}
+
+		declarations[i] = &genai.FunctionDeclaration{
+			Name:        info.Name,
+			Description: info.Description,
+			Parameters: &genai.Schema{
+				Type:       genai.TypeObject,
+				Properties: properties,
+				Required:   info.Required,
+			},
+		}
+	}
+
+	return declarations
+}
+
+// extractTokenUsage extracts token usage information from Gemini's response
+func (p *geminiProvider) extractTokenUsage(resp *genai.GenerateContentResponse) TokenUsage {
+	if resp == nil || resp.UsageMetadata == nil {
+		return TokenUsage{}
+	}
+
+	return TokenUsage{
+		InputTokens:         int64(resp.UsageMetadata.PromptTokenCount),
+		OutputTokens:        int64(resp.UsageMetadata.CandidatesTokenCount),
+		CacheCreationTokens: 0, // Not directly provided by Gemini
+		CacheReadTokens:     int64(resp.UsageMetadata.CachedContentTokenCount),
+	}
+}
+
+// SendMessages sends a batch of messages to Gemini and returns the response
+func (p *geminiProvider) SendMessages(ctx context.Context, messages []message.Message, tools []tools.BaseTool) (*ProviderResponse, error) {
+	// Create a generative model
+	model := p.client.GenerativeModel(p.model.APIModel)
+	model.SetMaxOutputTokens(p.maxTokens)
+
+	// Set system instruction
+	model.SystemInstruction = genai.NewUserContent(genai.Text(p.systemMessage))
+
+	// Set up tools if provided
+	if len(tools) > 0 {
+		declarations := p.convertToolsToGeminiFunctionDeclarations(tools)
+		model.Tools = []*genai.Tool{{FunctionDeclarations: declarations}}
+	}
+
+	// Create chat session and set history
+	chat := model.StartChat()
+	chat.History = p.convertToGeminiHistory(messages[:len(messages)-1]) // Exclude last message
+
+	// Get the most recent user message
+	var lastUserMsg message.Message
+	for i := len(messages) - 1; i >= 0; i-- {
+		if messages[i].Role == message.User {
+			lastUserMsg = messages[i]
+			break
+		}
+	}
+
+	// Send the message
+	resp, err := chat.SendMessage(ctx, genai.Text(lastUserMsg.Content))
+	if err != nil {
+		return nil, err
+	}
+
+	// Process the response
+	var content string
+	var toolCalls []message.ToolCall
+
+	if len(resp.Candidates) > 0 && resp.Candidates[0].Content != nil {
+		for _, part := range resp.Candidates[0].Content.Parts {
+			switch p := part.(type) {
+			case genai.Text:
+				content = string(p)
+			case genai.FunctionCall:
+				id := "call_" + uuid.New().String()
+				args, _ := json.Marshal(p.Args)
+				toolCalls = append(toolCalls, message.ToolCall{
+					ID:    id,
+					Name:  p.Name,
+					Input: string(args),
+					Type:  "function",
+				})
+			}
+		}
+	}
+
+	// Extract token usage
+	tokenUsage := p.extractTokenUsage(resp)
+
+	return &ProviderResponse{
+		Content:   content,
+		ToolCalls: toolCalls,
+		Usage:     tokenUsage,
+	}, nil
+}
+
+// StreamResponse streams the response from Gemini
+func (p *geminiProvider) StreamResponse(ctx context.Context, messages []message.Message, tools []tools.BaseTool) (<-chan ProviderEvent, error) {
+	// Create a generative model
+	model := p.client.GenerativeModel(p.model.APIModel)
+	model.SetMaxOutputTokens(p.maxTokens)
+
+	// Set system instruction
+	model.SystemInstruction = genai.NewUserContent(genai.Text(p.systemMessage))
+
+	// Set up tools if provided
+	if len(tools) > 0 {
+		declarations := p.convertToolsToGeminiFunctionDeclarations(tools)
+		for _, declaration := range declarations {
+			model.Tools = append(model.Tools, &genai.Tool{FunctionDeclarations: []*genai.FunctionDeclaration{declaration}})
+		}
+	}
+
+	// Create chat session and set history
+	chat := model.StartChat()
+	chat.History = p.convertToGeminiHistory(messages[:len(messages)-1]) // Exclude last message
+
+	lastUserMsg := messages[len(messages)-1]
+
+	// Start streaming
+	iter := chat.SendMessageStream(ctx, genai.Text(lastUserMsg.Content))
+
+	eventChan := make(chan ProviderEvent)
+
+	go func() {
+		defer close(eventChan)
+
+		var finalResp *genai.GenerateContentResponse
+		currentContent := ""
+		toolCalls := []message.ToolCall{}
+
+		for {
+			resp, err := iter.Next()
+			if err == iterator.Done {
+				break
+			}
+			if err != nil {
+				var apiErr *googleapi.Error
+				if errors.As(err, &apiErr) {
+					log.Printf("%s", apiErr.Body)
+				}
+				eventChan <- ProviderEvent{
+					Type:  EventError,
+					Error: err,
+				}
+				return
+			}
+
+			finalResp = resp
+
+			if len(resp.Candidates) > 0 && resp.Candidates[0].Content != nil {
+				for _, part := range resp.Candidates[0].Content.Parts {
+					switch p := part.(type) {
+					case genai.Text:
+						newText := string(p)
+						eventChan <- ProviderEvent{
+							Type:    EventContentDelta,
+							Content: newText,
+						}
+						currentContent += newText
+					case genai.FunctionCall:
+						// For function calls, we assume they come complete, not streamed in parts
+						id := "call_" + uuid.New().String()
+						args, _ := json.Marshal(p.Args)
+						newCall := message.ToolCall{
+							ID:    id,
+							Name:  p.Name,
+							Input: string(args),
+							Type:  "function",
+						}
+
+						// Check if this is a new tool call
+						isNew := true
+						for _, existing := range toolCalls {
+							if existing.Name == newCall.Name && existing.Input == newCall.Input {
+								isNew = false
+								break
+							}
+						}
+
+						if isNew {
+							toolCalls = append(toolCalls, newCall)
+						}
+					}
+				}
+			}
+		}
+
+		// Extract token usage from the final response
+		tokenUsage := p.extractTokenUsage(finalResp)
+
+		eventChan <- ProviderEvent{
+			Type: EventComplete,
+			Response: &ProviderResponse{
+				Content:   currentContent,
+				ToolCalls: toolCalls,
+				Usage:     tokenUsage,
+			},
+		}
+	}()
+
+	return eventChan, nil
+}
+
+// Helper function to parse JSON string into map
+func parseJsonToMap(jsonStr string) (map[string]interface{}, error) {
+	var result map[string]interface{}
+	err := json.Unmarshal([]byte(jsonStr), &result)
+	return result, err
+}
diff --git a/internal/llm/provider/openai.go b/internal/llm/provider/openai.go
new file mode 100644
index 000000000..d86a58690
--- /dev/null
+++ b/internal/llm/provider/openai.go
@@ -0,0 +1,278 @@
+package provider
+
+import (
+	"context"
+	"errors"
+
+	"github.com/kujtimiihoxha/termai/internal/llm/models"
+	"github.com/kujtimiihoxha/termai/internal/llm/tools"
+	"github.com/kujtimiihoxha/termai/internal/message"
+	"github.com/openai/openai-go"
+	"github.com/openai/openai-go/option"
+)
+
+type openaiProvider struct {
+	client        openai.Client
+	model         models.Model
+	maxTokens     int64
+	baseURL       string
+	apiKey        string
+	systemMessage string
+}
+
+type OpenAIOption func(*openaiProvider)
+
+func NewOpenAIProvider(opts ...OpenAIOption) (Provider, error) {
+	provider := &openaiProvider{
+		maxTokens: 5000,
+	}
+
+	for _, opt := range opts {
+		opt(provider)
+	}
+
+	clientOpts := []option.RequestOption{
+		option.WithAPIKey(provider.apiKey),
+	}
+	if provider.baseURL != "" {
+		clientOpts = append(clientOpts, option.WithBaseURL(provider.baseURL))
+	}
+
+	provider.client = openai.NewClient(clientOpts...)
+	if provider.systemMessage == "" {
+		return nil, errors.New("system message is required")
+	}
+
+	return provider, nil
+}
+
+func WithOpenAISystemMessage(message string) OpenAIOption {
+	return func(p *openaiProvider) {
+		p.systemMessage = message
+	}
+}
+
+func WithOpenAIMaxTokens(maxTokens int64) OpenAIOption {
+	return func(p *openaiProvider) {
+		p.maxTokens = maxTokens
+	}
+}
+
+func WithOpenAIModel(model models.Model) OpenAIOption {
+	return func(p *openaiProvider) {
+		p.model = model
+	}
+}
+
+func WithOpenAIBaseURL(baseURL string) OpenAIOption {
+	return func(p *openaiProvider) {
+		p.baseURL = baseURL
+	}
+}
+
+func WithOpenAIKey(apiKey string) OpenAIOption {
+	return func(p *openaiProvider) {
+		p.apiKey = apiKey
+	}
+}
+
+func (p *openaiProvider) convertToOpenAIMessages(messages []message.Message) []openai.ChatCompletionMessageParamUnion {
+	var chatMessages []openai.ChatCompletionMessageParamUnion
+
+	chatMessages = append(chatMessages, openai.SystemMessage(p.systemMessage))
+
+	for _, msg := range messages {
+		switch msg.Role {
+		case message.User:
+			chatMessages = append(chatMessages, openai.UserMessage(msg.Content))
+
+		case message.Assistant:
+			assistantMsg := openai.ChatCompletionAssistantMessageParam{
+				Role: "assistant",
+			}
+
+			if msg.Content != "" {
+				assistantMsg.Content = openai.ChatCompletionAssistantMessageParamContentUnion{
+					OfString: openai.String(msg.Content),
+				}
+			}
+
+			if len(msg.ToolCalls) > 0 {
+				assistantMsg.ToolCalls = make([]openai.ChatCompletionMessageToolCallParam, len(msg.ToolCalls))
+				for i, call := range msg.ToolCalls {
+					assistantMsg.ToolCalls[i] = openai.ChatCompletionMessageToolCallParam{
+						ID:   call.ID,
+						Type: "function",
+						Function: openai.ChatCompletionMessageToolCallFunctionParam{
+							Name:      call.Name,
+							Arguments: call.Input,
+						},
+					}
+				}
+			}
+
+			chatMessages = append(chatMessages, openai.ChatCompletionMessageParamUnion{
+				OfAssistant: &assistantMsg,
+			})
+
+		case message.Tool:
+			for _, result := range msg.ToolResults {
+				chatMessages = append(chatMessages,
+					openai.ToolMessage(result.Content, result.ToolCallID),
+				)
+			}
+		}
+	}
+
+	return chatMessages
+}
+
+func (p *openaiProvider) convertToOpenAITools(tools []tools.BaseTool) []openai.ChatCompletionToolParam {
+	openaiTools := make([]openai.ChatCompletionToolParam, len(tools))
+
+	for i, tool := range tools {
+		info := tool.Info()
+		openaiTools[i] = openai.ChatCompletionToolParam{
+			Function: openai.FunctionDefinitionParam{
+				Name:        info.Name,
+				Description: openai.String(info.Description),
+				Parameters: openai.FunctionParameters{
+					"type":       "object",
+					"properties": info.Parameters,
+					"required":   info.Required,
+				},
+			},
+		}
+	}
+
+	return openaiTools
+}
+
+func (p *openaiProvider) extractTokenUsage(usage openai.CompletionUsage) TokenUsage {
+	cachedTokens := int64(0)
+
+	cachedTokens = usage.PromptTokensDetails.CachedTokens
+	inputTokens := usage.PromptTokens - cachedTokens
+
+	return TokenUsage{
+		InputTokens:         inputTokens,
+		OutputTokens:        usage.CompletionTokens,
+		CacheCreationTokens: 0, // OpenAI doesn't provide this directly
+		CacheReadTokens:     cachedTokens,
+	}
+}
+
+func (p *openaiProvider) SendMessages(ctx context.Context, messages []message.Message, tools []tools.BaseTool) (*ProviderResponse, error) {
+	chatMessages := p.convertToOpenAIMessages(messages)
+	openaiTools := p.convertToOpenAITools(tools)
+
+	params := openai.ChatCompletionNewParams{
+		Model:     openai.ChatModel(p.model.APIModel),
+		Messages:  chatMessages,
+		MaxTokens: openai.Int(p.maxTokens),
+		Tools:     openaiTools,
+	}
+
+	response, err := p.client.Chat.Completions.New(ctx, params)
+	if err != nil {
+		return nil, err
+	}
+
+	content := ""
+	if response.Choices[0].Message.Content != "" {
+		content = response.Choices[0].Message.Content
+	}
+
+	var toolCalls []message.ToolCall
+	if len(response.Choices[0].Message.ToolCalls) > 0 {
+		toolCalls = make([]message.ToolCall, len(response.Choices[0].Message.ToolCalls))
+		for i, call := range response.Choices[0].Message.ToolCalls {
+			toolCalls[i] = message.ToolCall{
+				ID:    call.ID,
+				Name:  call.Function.Name,
+				Input: call.Function.Arguments,
+				Type:  "function",
+			}
+		}
+	}
+
+	tokenUsage := p.extractTokenUsage(response.Usage)
+
+	return &ProviderResponse{
+		Content:   content,
+		ToolCalls: toolCalls,
+		Usage:     tokenUsage,
+	}, nil
+}
+
+func (p *openaiProvider) StreamResponse(ctx context.Context, messages []message.Message, tools []tools.BaseTool) (<-chan ProviderEvent, error) {
+	chatMessages := p.convertToOpenAIMessages(messages)
+	openaiTools := p.convertToOpenAITools(tools)
+
+	params := openai.ChatCompletionNewParams{
+		Model:     openai.ChatModel(p.model.APIModel),
+		Messages:  chatMessages,
+		MaxTokens: openai.Int(p.maxTokens),
+		Tools:     openaiTools,
+		StreamOptions: openai.ChatCompletionStreamOptionsParam{
+			IncludeUsage: openai.Bool(true),
+		},
+	}
+
+	stream := p.client.Chat.Completions.NewStreaming(ctx, params)
+
+	eventChan := make(chan ProviderEvent)
+
+	toolCalls := make([]message.ToolCall, 0)
+	go func() {
+		defer close(eventChan)
+
+		acc := openai.ChatCompletionAccumulator{}
+		currentContent := ""
+
+		for stream.Next() {
+			chunk := stream.Current()
+			acc.AddChunk(chunk)
+
+			if tool, ok := acc.JustFinishedToolCall(); ok {
+				toolCalls = append(toolCalls, message.ToolCall{
+					ID:    tool.Id,
+					Name:  tool.Name,
+					Input: tool.Arguments,
+					Type:  "function",
+				})
+			}
+
+			for _, choice := range chunk.Choices {
+				if choice.Delta.Content != "" {
+					eventChan <- ProviderEvent{
+						Type:    EventContentDelta,
+						Content: choice.Delta.Content,
+					}
+					currentContent += choice.Delta.Content
+				}
+			}
+		}
+
+		if err := stream.Err(); err != nil {
+			eventChan <- ProviderEvent{
+				Type:  EventError,
+				Error: err,
+			}
+			return
+		}
+
+		tokenUsage := p.extractTokenUsage(acc.Usage)
+
+		eventChan <- ProviderEvent{
+			Type: EventComplete,
+			Response: &ProviderResponse{
+				Content:   currentContent,
+				ToolCalls: toolCalls,
+				Usage:     tokenUsage,
+			},
+		}
+	}()
+
+	return eventChan, nil
+}
diff --git a/internal/llm/provider/provider.go b/internal/llm/provider/provider.go
new file mode 100644
index 000000000..9ac1def37
--- /dev/null
+++ b/internal/llm/provider/provider.go
@@ -0,0 +1,48 @@
+package provider
+
+import (
+	"context"
+
+	"github.com/kujtimiihoxha/termai/internal/llm/tools"
+	"github.com/kujtimiihoxha/termai/internal/message"
+)
+
+// EventType represents the type of streaming event
+type EventType string
+
+const (
+	EventContentStart  EventType = "content_start"
+	EventContentDelta  EventType = "content_delta"
+	EventThinkingDelta EventType = "thinking_delta"
+	EventContentStop   EventType = "content_stop"
+	EventComplete      EventType = "complete"
+	EventError         EventType = "error"
+)
+
+type TokenUsage struct {
+	InputTokens         int64
+	OutputTokens        int64
+	CacheCreationTokens int64
+	CacheReadTokens     int64
+}
+
+type ProviderResponse struct {
+	Content   string
+	ToolCalls []message.ToolCall
+	Usage     TokenUsage
+}
+
+type ProviderEvent struct {
+	Type     EventType
+	Content  string
+	Thinking string
+	ToolCall *message.ToolCall
+	Error    error
+	Response *ProviderResponse
+}
+
+type Provider interface {
+	SendMessages(ctx context.Context, messages []message.Message, tools []tools.BaseTool) (*ProviderResponse, error)
+
+	StreamResponse(ctx context.Context, messages []message.Message, tools []tools.BaseTool) (<-chan ProviderEvent, error)
+}