Image input

internal/claude/session.go

@@ -5,6 +5,7 @@ import (
 	"bytes"
 	"context"
 	"crypto/rand"
+	"encoding/base64"
 	"encoding/hex"
 	"encoding/json"
 	"errors"
@@ -359,10 +360,34 @@ func (s *Session) ID() string {
 	return s.id
 }
 
-func (s *Session) Send(text string) error {
+type Image struct {
+	MediaType string
+	Data      []byte
+}
+
+func (s *Session) Send(text string, images ...Image) error {
+	var content any = text
+	if len(images) > 0 {
+		blocks := []map[string]any{}
+		for _, img := range images {
+			blocks = append(blocks, map[string]any{
+				"type": "image",
+				"source": map[string]any{
+					"type":       "base64",
+					"media_type": img.MediaType,
+					"data":       base64.StdEncoding.EncodeToString(img.Data),
+				},
+			})
+		}
+		if text != "" {
+			blocks = append(blocks, map[string]any{"type": "text", "text": text})
+		}
+		content = blocks
+	}
+
 	return s.write(map[string]any{
 		"type":               "user",
-		"message":            map[string]any{"role": "user", "content": text},
+		"message":            map[string]any{"role": "user", "content": content},
 		"parent_tool_use_id": nil,
 		"session_id":         s.ID(),
 	})

internal/claude/session_test.go

@@ -1,8 +1,13 @@
 package claude_test
 
 import (
+	"bytes"
 	"context"
 	"encoding/json"
+	"image"
+	"image/color"
+	"image/draw"
+	"image/png"
 	"os"
 	"os/exec"
 	"path/filepath"
@@ -198,6 +203,30 @@ func TestInterrupt(t *testing.T) {
 	}
 }
 
+func TestImageInput(t *testing.T) {
+	t.Parallel()
+	ctx := testContext(t)
+	s := start(t, ctx, claude.Options{Cwd: scratchRepo(t)})
+
+	img := image.NewRGBA(image.Rect(0, 0, 64, 64))
+	draw.Draw(img, img.Bounds(), &image.Uniform{C: color.RGBA{R: 255, A: 255}}, image.Point{}, draw.Src)
+	buf := bytes.Buffer{}
+	err := png.Encode(&buf, img)
+	if err != nil {
+		t.Fatalf("encode png: %v", err)
+	}
+
+	err = s.Send("Without using any tools, what single colour fills the attached image? Reply with just the colour name.", claude.Image{MediaType: "image/png", Data: buf.Bytes()})
+	if err != nil {
+		t.Fatalf("send: %v", err)
+	}
+	text := waitResult(t, ctx, s, nil)
+
+	if !strings.Contains(strings.ToLower(text), "red") {
+		t.Fatalf("model didn't see the image; assistant said %q", text)
+	}
+}
+
 func TestResume(t *testing.T) {
 	t.Parallel()
 	ctx := testContext(t)