free-ide-proxy/cmd/mcp-vision/main.go

package main

import (
	"bufio"
	"bytes"
	"encoding/base64"
	"encoding/json"
	"fmt"
	"io"
	"net/http"
	"os"
	"path/filepath"
	"strings"
)

const (
	defaultAPIKey = "opx-57ffcb013651f7acae558cd9ab094dbd719fdec980a6ad53"
	proxyBaseURL  = "http://127.0.0.1:6446/v1"
)

func main() {
	debug := os.Getenv("MCP_DEBUG") != ""
	logf := func(f string, a ...any) {
		if debug {
			fmt.Fprintf(os.Stderr, f+"\n", a...)
		}
	}

	reader := bufio.NewReader(os.Stdin)
	buf := make([]byte, 0, 64*1024)

	for {
		var msg map[string]any

		first, err := reader.ReadString('\n')
		if err != nil {
			logf("stdin EOF: %v", err)
			return
		}
		logf("RAW stdin first line: %q", first)

		if strings.HasPrefix(first, "Content-Length: ") {
			var contentLen int
			fmt.Sscanf(first, "Content-Length: %d", &contentLen)
			logf("mode=Content-Length len=%d", contentLen)

			emptyLine, err := reader.ReadString('\n')
			if err != nil {
				logf("err reading empty line: %v", err)
				return
			}
			logf("empty line: %q", emptyLine)

			buf = buf[:contentLen]
			if _, err := io.ReadFull(reader, buf); err != nil {
				logf("err reading body: %v", err)
				return
			}
			logf("body: %s", string(buf))
			json.Unmarshal(buf, &msg)
		} else if len(first) > 0 && first[0] == '{' {
			// Raw JSON line mode (like engram)
			logf("mode=rawJSON")
			json.Unmarshal([]byte(strings.TrimRight(first, "\r\n")), &msg)
		} else {
			logf("unrecognized input, skipping: %q", first)
			continue
		}

		if msg == nil {
			logf("msg is nil, skipping")
			continue
		}

		method, _ := msg["method"].(string)
		id, _ := msg["id"]

		switch method {
		case "initialize":
			respond(id, map[string]any{
				"protocolVersion": "2025-03-26",
				"capabilities": map[string]any{
					"tools": map[string]any{},
				},
				"serverInfo": map[string]any{
					"name":    "mcp-vision",
					"version": "1.0.0",
				},
			})

		case "tools/list":
			respond(id, map[string]any{
				"tools": []map[string]any{
					{
						"name":        "read_image",
						"description": "Describe an image file using AI vision. Sends only the image + prompt, no conversation context.",
						"inputSchema": map[string]any{
							"type": "object",
							"properties": map[string]any{
								"path": map[string]any{
									"type":        "string",
									"description": "Absolute path to the image file (jpg, png, webp)",
								},
								"prompt": map[string]any{
									"type":        "string",
									"description": "Optional custom question about the image",
								},
							},
							"required": []string{"path"},
						},
					},
				},
			})

		case "tools/call":
			params, _ := msg["params"].(map[string]any)
			name, _ := params["name"].(string)
			args, _ := params["arguments"].(map[string]any)

			if name != "read_image" {
				respondError(id, -32601, fmt.Sprintf("Unknown tool: %s", name))
				continue
			}

			path, _ := args["path"].(string)
			prompt, _ := args["prompt"].(string)
			if prompt == "" {
				prompt = "Describe esta imagen en detalle."
			}

			result, err := readImage(path, prompt)
			if err != nil {
				respondError(id, -32000, err.Error())
				continue
			}

			respond(id, map[string]any{
				"content": []map[string]any{
					{"type": "text", "text": result},
				},
			})

		case "notifications/initialized":
		case "notifications/cancelled":
		}
	}
}

func readImage(path, prompt string) (string, error) {
	ext := strings.ToLower(filepath.Ext(path))
	var mediaType string
	switch ext {
	case ".jpg", ".jpeg":
		mediaType = "image/jpeg"
	case ".png":
		mediaType = "image/png"
	case ".webp":
		mediaType = "image/webp"
	default:
		return "", fmt.Errorf("unsupported format: %s (use jpg, png, or webp)", ext)
	}

	data, err := os.ReadFile(path)
	if err != nil {
		return "", fmt.Errorf("cannot read file: %w", err)
	}

	b64 := base64.StdEncoding.EncodeToString(data)

	reqBody := map[string]any{
		"model": "mimo-v2.5-free",
		"messages": []map[string]any{
			{
				"role": "user",
				"content": []map[string]any{
					{"type": "text", "text": prompt},
					{"type": "image_url", "image_url": map[string]string{
						"url": fmt.Sprintf("data:%s;base64,%s", mediaType, b64),
					}},
				},
			},
		},
		"stream": false,
	}

	body, _ := json.Marshal(reqBody)

	httpReq, err := http.NewRequest("POST", proxyBaseURL+"/chat/completions", bytes.NewReader(body))
	if err != nil {
		return "", fmt.Errorf("request error: %w", err)
	}

	apiKey := os.Getenv("VISION_API_KEY")
	if apiKey == "" {
		apiKey = defaultAPIKey
	}
	// using proxy at proxyBaseURL

	httpReq.Header.Set("Content-Type", "application/json")
	httpReq.Header.Set("Authorization", "Bearer "+apiKey)
	httpReq.Header.Set("User-Agent", "mcp-vision/1.0")
	httpReq.Header.Set("x-opencode-client", "mcp-vision")
	httpReq.Header.Set("x-opencode-project", "vision")
	httpReq.Header.Set("x-opencode-request", fmt.Sprintf("req_%d", len(data)))
	httpReq.Header.Set("x-opencode-session", fmt.Sprintf("ses_%d", len(data)%1000000))

	client := &http.Client{Timeout: 2 * 60}
	resp, err := client.Do(httpReq)
	if err != nil {
		return "", fmt.Errorf("api error: %w", err)
	}
	defer resp.Body.Close()

	respData, _ := io.ReadAll(resp.Body)
	if resp.StatusCode != 200 {
		return "", fmt.Errorf("api returned %d: %s", resp.StatusCode, string(respData))
	}

	var result struct {
		Choices []struct {
			Message struct {
				Content any `json:"content"`
			} `json:"message"`
		} `json:"choices"`
	}
	if err := json.Unmarshal(respData, &result); err != nil {
		return "", fmt.Errorf("parse error: %w", err)
	}

	if len(result.Choices) == 0 {
		return "", fmt.Errorf("empty response")
	}

	switch c := result.Choices[0].Message.Content.(type) {
	case string:
		return c, nil
	default:
		b, _ := json.MarshalIndent(c, "", "  ")
		return string(b), nil
	}
}

func respond(id any, result any) {
	resp := map[string]any{
		"jsonrpc": "2.0",
		"id":      id,
		"result":  result,
	}
	writeJSON(resp)
}

func respondError(id any, code int, message string) {
	resp := map[string]any{
		"jsonrpc": "2.0",
		"id":      id,
		"error": map[string]any{
			"code":    code,
			"message": message,
		},
	}
	writeJSON(resp)
}

func writeJSON(v any) {
	data, _ := json.Marshal(v)
	// Claude Code accepts newline-delimited JSON (same format as engram)
	os.Stdout.Write(data)
	os.Stdout.Write([]byte{'\n'})
}