Add Claude as a new AI provider

config/default.yaml

@@ -25,6 +25,8 @@ providers:
   vertexai:
     project: your-gcp-project
     location: us-central1
+  claude:
+    api_key: your-claude-api-key
 
 scheduler:
   sync_batch_job_status_cron: "* * * * *"

go.mod

@@ -3,6 +3,7 @@ module github.com/ideagate/aigateway-core
 go 1.25.8
 
 require (
+	github.com/anthropics/anthropic-sdk-go v1.33.0
 	github.com/redis/go-redis/v9 v9.18.0
 	github.com/robfig/cron/v3 v3.0.1
 	github.com/spf13/cast v1.7.1
@@ -49,6 +50,10 @@ require (
 	github.com/spf13/pflag v1.0.6 // indirect
 	github.com/stretchr/objx v0.5.2 // indirect
 	github.com/subosito/gotenv v1.6.0 // indirect
+	github.com/tidwall/gjson v1.18.0 // indirect
+	github.com/tidwall/match v1.1.1 // indirect
+	github.com/tidwall/pretty v1.2.1 // indirect
+	github.com/tidwall/sjson v1.2.5 // indirect
 	go.opentelemetry.io/auto/sdk v1.2.1 // indirect
 	go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.54.0 // indirect
 	go.opentelemetry.io/otel v1.39.0 // indirect

go.sum

@@ -6,6 +6,8 @@ cloud.google.com/go/compute/metadata v0.9.0 h1:pDUj4QMoPejqq20dK0Pg2N4yG9zIkYGdB
 cloud.google.com/go/compute/metadata v0.9.0/go.mod h1:E0bWwX5wTnLPedCKqk3pJmVgCBSM6qQI1yTBdEb3C10=
 filippo.io/edwards25519 v1.1.0 h1:FNf4tywRC1HmFuKW5xopWpigGjJKiJSV0Cqo0cJWDaA=
 filippo.io/edwards25519 v1.1.0/go.mod h1:BxyFTGdWcka3PhytdK4V28tE5sGfRvvvRV7EaN4VDT4=
+github.com/anthropics/anthropic-sdk-go v1.33.0 h1:YlRqiI+PjULBA8NoeeJmkXtFzTmjZZA7oFvpZ4FU7eU=
+github.com/anthropics/anthropic-sdk-go v1.33.0/go.mod h1:dSIO7kSrOI7MA4fE6RRVaw8tyWP7HNQU5/H/KS4cax8=
 github.com/bsm/ginkgo/v2 v2.12.0 h1:Ny8MWAHyOepLGlLKYmXG4IEkioBysk6GpaRTLC8zwWs=
 github.com/bsm/ginkgo/v2 v2.12.0/go.mod h1:SwYbGRRDovPVboqFv0tPTcG1sN61LM1Z4ARdbAV9g4c=
 github.com/bsm/gomega v1.27.10 h1:yeMWxP2pV2fG3FgAODIY8EiRE3dy0aeFYt4l7wh6yKA=
@@ -17,6 +19,8 @@ github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f h1:lO4WD4F/rVNCu3HqELle0jiPLLBs70cWOduZpkS1E78=
 github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f/go.mod h1:cuUVRXasLTGF7a8hSLbxyZXjz+1KgoB3wDUb6vlszIc=
+github.com/dnaeon/go-vcr v1.2.0 h1:zHCHvJYTMh1N7xnV7zf1m1GPBF9Ad0Jk/whtQ1663qI=
+github.com/dnaeon/go-vcr v1.2.0/go.mod h1:R4UdLID7HZT3taECzJs4YgbbH6PIGXB6W/sc5OLb6RQ=
 github.com/felixge/httpsnoop v1.0.4 h1:NFTV2Zj1bL4mc9sqWACXbQFVBBg2W3GPvqp8/ESS2Wg=
 github.com/felixge/httpsnoop v1.0.4/go.mod h1:m8KPJKqk1gH5J9DgRY2ASl2lWCfGKXixSwevea8zH2U=
 github.com/frankban/quicktest v1.14.6 h1:7Xjx+VpznH+oBnejlPUj8oUpdxnVs4f8XU8WnHkI4W8=
@@ -95,6 +99,16 @@ github.com/stretchr/testify v1.11.1 h1:7s2iGBzp5EwR7/aIZr8ao5+dra3wiQyKjjFuvgVKu
 github.com/stretchr/testify v1.11.1/go.mod h1:wZwfW3scLgRK+23gO65QZefKpKQRnfz6sD981Nm4B6U=
 github.com/subosito/gotenv v1.6.0 h1:9NlTDc1FTs4qu0DDq7AEtTPNw6SVm7uBMsUCUjABIf8=
 github.com/subosito/gotenv v1.6.0/go.mod h1:Dk4QP5c2W3ibzajGcXpNraDfq2IrhjMIvMSWPKKo0FU=
+github.com/tidwall/gjson v1.14.2/go.mod h1:/wbyibRr2FHMks5tjHJ5F8dMZh3AcwJEMf5vlfC0lxk=
+github.com/tidwall/gjson v1.18.0 h1:FIDeeyB800efLX89e5a8Y0BNH+LOngJyGrIWxG2FKQY=
+github.com/tidwall/gjson v1.18.0/go.mod h1:/wbyibRr2FHMks5tjHJ5F8dMZh3AcwJEMf5vlfC0lxk=
+github.com/tidwall/match v1.1.1 h1:+Ho715JplO36QYgwN9PGYNhgZvoUSc9X2c80KVTi+GA=
+github.com/tidwall/match v1.1.1/go.mod h1:eRSPERbgtNPcGhD8UCthc6PmLEQXEWd3PRB5JTxsfmM=
+github.com/tidwall/pretty v1.2.0/go.mod h1:ITEVvHYasfjBbM0u2Pg8T2nJnzm8xPwvNhhsoaGGjNU=
+github.com/tidwall/pretty v1.2.1 h1:qjsOFOWWQl+N3RsoF5/ssm1pHmJJwhjlSbZ51I6wMl4=
+github.com/tidwall/pretty v1.2.1/go.mod h1:ITEVvHYasfjBbM0u2Pg8T2nJnzm8xPwvNhhsoaGGjNU=
+github.com/tidwall/sjson v1.2.5 h1:kLy8mja+1c9jlljvWTlSazM7cKDRfJuR/bOJhcY5NcY=
+github.com/tidwall/sjson v1.2.5/go.mod h1:Fvgq9kS/6ociJEDnK0Fk1cpYF4FIW6ZF7LAe+6jwd28=
 github.com/zeebo/xxh3 v1.0.2 h1:xZmwmqxHZA8AI603jOQ0tMqmBr9lPeFwGg6d+xy9DC0=
 github.com/zeebo/xxh3 v1.0.2/go.mod h1:5NWz9Sef7zIDm2JHfFlcQvNekmcEl9ekUZQQKCYaDcA=
 go.opentelemetry.io/auto/sdk v1.2.1 h1:jXsnJ4Lmnqd11kwkBV2LgLoFMZKizbCi5fNZ/ipaZ64=
@@ -138,6 +152,8 @@ google.golang.org/protobuf v1.36.10/go.mod h1:HTf+CrKn2C3g5S8VImy6tdcUvCska2kB7j
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
 gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c/go.mod h1:JHkPIbrfpd72SG/EVd6muEfDQjcINNoR0C8j2r3qZ4Q=
+gopkg.in/yaml.v2 v2.2.8 h1:obN1ZagJSUGI0Ek/LBmuj4SNLPfIny3KsKFopxRdj10=
+gopkg.in/yaml.v2 v2.2.8/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
 gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
 gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
 gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=

internal/aigateway/providers/claude.go

@@ -0,0 +1,183 @@
+package providers
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"fmt"
+
+	"github.com/anthropics/anthropic-sdk-go"
+	"github.com/anthropics/anthropic-sdk-go/option"
+	"github.com/anthropics/anthropic-sdk-go/packages/param"
+	aigatewayv1 "github.com/ideagate/aigateway-core/gen/aigateway/v1"
+	"github.com/ideagate/aigateway-core/internal/aigateway/models"
+	models2 "github.com/ideagate/aigateway-core/models"
+)
+
+const claudeDefaultMaxTokens = 4096
+
+type ClaudeProvider struct {
+	client anthropic.Client
+}
+
+func newClaudeProvider(apiKey string) (*ClaudeProvider, error) {
+	if apiKey == "" {
+		return nil, fmt.Errorf("claude api key is required")
+	}
+
+	return &ClaudeProvider{
+		client: anthropic.NewClient(option.WithAPIKey(apiKey)),
+	}, nil
+}
+
+func (c *ClaudeProvider) Name() string {
+	return "claude"
+}
+
+func (c *ClaudeProvider) ChatCompletion(ctx context.Context, request *models2.ChatCompletionRequest) (*models2.ChatCompletionResponse, error) {
+	params := anthropic.MessageNewParams{
+		Model:     anthropic.Model(request.Model),
+		MaxTokens: claudeDefaultMaxTokens,
+		Messages: []anthropic.MessageParam{
+			anthropic.NewUserMessage(anthropic.NewTextBlock(request.Content.Text)),
+		},
+	}
+
+	if request.Temperature != 0 {
+		params.Temperature = param.NewOpt(float64(request.Temperature))
+	}
+
+	if request.SystemInstruction.Text != "" {
+		params.System = []anthropic.TextBlockParam{
+			{Text: request.SystemInstruction.Text},
+		}
+	}
+
+	msg, err := c.client.Messages.New(ctx, params)
+	if err != nil {
+		return nil, fmt.Errorf("create message: %w", err)
+	}
+
+	text := extractTextFromContent(msg.Content)
+
+	return &models2.ChatCompletionResponse{
+		Content: models2.Content{
+			Text: text,
+		},
+	}, nil
+}
+
+func (c *ClaudeProvider) SubmitBatchJob(ctx context.Context, requests []*aigatewayv1.SubmitBulkChatCompletionsRequest) (*aigatewayv1.SubmitBulkChatCompletionsResponse, error) {
+	if len(requests) == 0 {
+		return nil, errors.New("no requests provided")
+	}
+
+	batchRequests := make([]anthropic.MessageBatchNewParamsRequest, len(requests))
+	for i, request := range requests {
+		reqParams := anthropic.MessageBatchNewParamsRequestParams{
+			Model:     anthropic.Model(request.GetModel()),
+			MaxTokens: claudeDefaultMaxTokens,
+			Messages: []anthropic.MessageParam{
+				anthropic.NewUserMessage(anthropic.NewTextBlock(request.GetContent().GetText())),
+			},
+		}
+
+		if request.GetTemperature() != 0 {
+			reqParams.Temperature = param.NewOpt(float64(request.GetTemperature()))
+		}
+
+		if request.GetSystemInstruction().GetText() != "" {
+			reqParams.System = []anthropic.TextBlockParam{
+				{Text: request.GetSystemInstruction().GetText()},
+			}
+		}
+
+		batchRequests[i] = anthropic.MessageBatchNewParamsRequest{
+			CustomID: fmt.Sprintf("%d", i),
+			Params:   reqParams,
+		}
+	}
+
+	batch, err := c.client.Messages.Batches.New(ctx, anthropic.MessageBatchNewParams{
+		Requests: batchRequests,
+	})
+	if err != nil {
+		return nil, fmt.Errorf("failed to create batch job: %w", err)
+	}
+
+	return &aigatewayv1.SubmitBulkChatCompletionsResponse{
+		JobId: batch.ID,
+	}, nil
+}
+
+// GetBatchJobStatus polls the provider once and returns the current status.
+// On completion the full individual responses slice is JSON-serialised into
+// ResultsJSON so callers can persist it directly.
+func (c *ClaudeProvider) GetBatchJobStatus(ctx context.Context, referenceID string) (*BatchJobStatusResult, error) {
+	batch, err := c.client.Messages.Batches.Get(ctx, referenceID)
+	if err != nil {
+		return nil, fmt.Errorf("get batch job %s: %w", referenceID, err)
+	}
+
+	result := &BatchJobStatusResult{}
+
+	switch batch.ProcessingStatus {
+	case anthropic.MessageBatchProcessingStatusEnded:
+		// Collect all individual results by streaming the results file.
+		stream := c.client.Messages.Batches.ResultsStreaming(ctx, referenceID)
+		var responses []anthropic.MessageBatchIndividualResponse
+		for stream.Next() {
+			responses = append(responses, stream.Current())
+		}
+		if err := stream.Err(); err != nil {
+			return nil, fmt.Errorf("stream batch results %s: %w", referenceID, err)
+		}
+
+		// Determine overall status and aggregate token counts.
+		// Batches with at least one succeeded request are marked completed,
+		// matching the GoogleProvider behaviour for partial success. Callers
+		// can inspect the individual ResultsJSON entries for per-request errors.
+		succeeded := batch.RequestCounts.Succeeded
+		if succeeded == 0 {
+			result.Status = models.BatchJobStatusFailed
+		} else {
+			result.Status = models.BatchJobStatusCompleted
+		}
+
+		for _, r := range responses {
+			if r.Result.Type == "succeeded" {
+				usage := r.Result.Message.Usage
+				result.InputTokenCount += usage.InputTokens
+				result.OutputTokenCount += usage.OutputTokens
+				result.TotalTokenCount += usage.InputTokens + usage.OutputTokens
+			}
+		}
+
+		resultsJSON, err := json.Marshal(responses)
+		if err != nil {
+			return nil, fmt.Errorf("marshal batch results: %w", err)
+		}
+		result.ResultsJSON = resultsJSON
+
+	case anthropic.MessageBatchProcessingStatusCanceling:
+		result.Status = models.BatchJobStatusProcessing
+
+	default:
+		// in_progress — still in flight.
+		result.Status = models.BatchJobStatusProcessing
+	}
+
+	return result, nil
+}
+
+// extractTextFromContent returns the concatenated text from all text-type
+// content blocks in the message response.
+func extractTextFromContent(content []anthropic.ContentBlockUnion) string {
+	var text string
+	for _, block := range content {
+		if block.Type == "text" {
+			text += block.Text
+		}
+	}
+	return text
+}

internal/aigateway/providers/claude_test.go

@@ -0,0 +1,80 @@
+package providers
+
+import (
+	"testing"
+
+	"github.com/anthropics/anthropic-sdk-go"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestNewClaudeProvider(t *testing.T) {
+	t.Run("empty api key returns error", func(t *testing.T) {
+		_, err := newClaudeProvider("")
+		assert.Error(t, err)
+		assert.Contains(t, err.Error(), "api key is required")
+	})
+
+	t.Run("valid api key returns provider", func(t *testing.T) {
+		p, err := newClaudeProvider("test-api-key")
+		require.NoError(t, err)
+		assert.NotNil(t, p)
+		assert.Equal(t, "claude", p.Name())
+	})
+}
+
+func TestExtractTextFromContent(t *testing.T) {
+	tests := []struct {
+		name     string
+		content  []anthropic.ContentBlockUnion
+		wantText string
+	}{
+		{
+			name:     "nil slice",
+			content:  nil,
+			wantText: "",
+		},
+		{
+			name:     "empty slice",
+			content:  []anthropic.ContentBlockUnion{},
+			wantText: "",
+		},
+		{
+			name: "single text block",
+			content: []anthropic.ContentBlockUnion{
+				{Type: "text", Text: "Hello, world!"},
+			},
+			wantText: "Hello, world!",
+		},
+		{
+			name: "multiple text blocks concatenated",
+			content: []anthropic.ContentBlockUnion{
+				{Type: "text", Text: "Hello"},
+				{Type: "text", Text: ", world!"},
+			},
+			wantText: "Hello, world!",
+		},
+		{
+			name: "non-text blocks skipped",
+			content: []anthropic.ContentBlockUnion{
+				{Type: "thinking", Thinking: "some reasoning"},
+				{Type: "text", Text: "Answer"},
+			},
+			wantText: "Answer",
+		},
+		{
+			name: "no text blocks returns empty string",
+			content: []anthropic.ContentBlockUnion{
+				{Type: "thinking", Thinking: "some reasoning"},
+			},
+			wantText: "",
+		},
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			got := extractTextFromContent(tc.content)
+			assert.Equal(t, tc.wantText, got)
+		})
+	}
+}

internal/aigateway/providers/providers.go

@@ -33,3 +33,7 @@ type Provider interface {
 func New(geminiCfg platformconfig.GeminiConfig) (Provider, error) {
 	return newGoogleProvider(geminiCfg.APIKey)
 }
+
+func NewClaude(claudeCfg platformconfig.ClaudeConfig) (Provider, error) {
+	return newClaudeProvider(claudeCfg.APIKey)
+}

internal/platform/config/config.go

@@ -43,6 +43,7 @@ type RedisConfig struct {
 type ProvidersConfig struct {
 	Gemini   GeminiConfig   `mapstructure:"gemini"`
 	VertexAI VertexAIConfig `mapstructure:"vertexai"`
+	Claude   ClaudeConfig   `mapstructure:"claude"`
 }
 
 type GeminiConfig struct {
@@ -54,6 +55,10 @@ type VertexAIConfig struct {
 	Location string `mapstructure:"location"`
 }
 
+type ClaudeConfig struct {
+	APIKey string `mapstructure:"api_key"`
+}
+
 type SchedulerConfig struct {
 	SyncBatchJobStatusCron string `mapstructure:"sync_batch_job_status_cron"`
 }