feat: Add Telegram file/image support with vision, OCR, and document-to-context pipeline

Cargo.toml

@@ -55,5 +55,22 @@ sqlite-vec = "0.1"
 # Setup wizard web server (used only by src/bin/setup.rs)
 axum = "0.8"
 
+# OCR (pure Rust, neural-network based)
+ocrs = "0.12"
+rten = { version = "0.24", features = ["rten_format"] }
+
+# Image loading/processing
+image = { version = "0.25", default-features = false, features = ["jpeg", "png", "gif", "webp"] }
+
+# Document processing
+pdf-extract = "0.10"
+docx-rs = "0.4"
+
+# MIME type detection
+infer = "0.19"
+
+# Base64 for vision API content parts
+base64 = "0.22"
+
 [dev-dependencies]
 tempfile = "3"

config.example.toml

@@ -16,6 +16,7 @@ model = "moonshotai/kimi-k2.5"
 base_url = "https://openrouter.ai/api/v1"
 # Maximum tokens in response
 max_tokens = 4096
+# supports_vision = false  # Set to true if your model supports image inputs
 # System prompt for the AI assistant
 system_prompt = """You are a helpful AI assistant with access to tools. \
 Use the available tools to help the user with their tasks. \
@@ -63,6 +64,12 @@ directory = "skills"
 # model = "qwen/qwen3-embedding-8b"
 # dimensions = 1536
 
+# OCR configuration (optional)
+# Used to extract text from images when supports_vision = false.
+# Models are downloaded automatically on first use.
+# [ocr]
+# model_dir = "~/.cache/ocrs"  # Where OCR model files are cached (downloaded on first use)
+
 # MCP Server Configurations
 # Each [[mcp_servers]] block defines an MCP server to connect to
 # The bot will discover and register tools from each server