Plugins
KairosRoute includes built-in plugins that extend what your models can do. Plugins are exposed as tools your models can call during chat completions — no extra integration required.
Access tiers
Free plugins
Available to all users. Web search, Project Gutenberg, and code execution are enabled by default on every request.
Premium plugins
Require any positive credit balance. Image generation, PDF reader, and agent memory unlock when you add credits.
Enabling plugins
Pass the plugin IDs you want in the plugins array of your request body. Free plugins are always available; premium plugins require a paid account.
{
"model": "kr-auto",
"messages": [...],
"plugins": ["web-search", "image-gen", "pdf-reader", "memory"]
}Plugin reference
Web Search
FreeID: web-search
Gives models access to real-time web search results. The model can search the web and incorporate current information into its responses. Enabled by default.
Tools: search_web
Project Gutenberg
FreeID: gutenberg
Search and read full texts from the Project Gutenberg library of 70,000+ free ebooks. Models can look up passages, search for titles, and retrieve book content.
Tools: gutenberg_search, gutenberg_read
Code Execution
FreeID: code-exec
Sandboxed code execution in Python and JavaScript. Models can write and run code to perform calculations, data analysis, and generate outputs.
Tools: execute_code
Image Generation
PremiumID: image-gen
Generate images from text prompts using multiple providers. Supports DALL-E 3, Flux 1.1 Pro, and Nano Banana. Use auto mode to automatically select the cheapest available provider.
Provider pricing
Flux 1.1 Pro — $0.025/image
Nano Banana (Gemini) — $0.039/image
DALL-E 3 Standard — $0.040/image
DALL-E 3 HD — $0.080/image
Tools: generate_image (params: prompt, provider, size, quality, style)
PDF Reader
PremiumID: pdf-reader
Extract text, tables, and metadata from PDF documents. Supports URL-based and base64 uploads, page range selection, and table detection. Up to 50MB per document.
Tools: read_pdf (params: source, source_type, pages, max_chars, extract_tables), get_pdf_info
Agent Memory
PremiumID: memory
Persistent vector memory for AI agents. Store, search, and retrieve information across conversations using semantic similarity. Powered by pgvector with OpenAI embeddings. Supports namespaces, tags, and TTL-based expiration.
Example usage
// Agent stores a fact
memory_store({
content: "User prefers Python over JS",
namespace: "user-prefs",
tags: ["language", "preference"]
})
// Agent recalls later
memory_search({
query: "what language does the user prefer",
namespace: "user-prefs",
limit: 3
})Tools: memory_store, memory_search, memory_delete
Try plugins today
Free plugins are available immediately. Add credits to unlock premium plugins.