API Documentation

Generis AI API

Complete reference for integrating Generis AI into your applications.

📑 Contents

🚀 Getting Started

All API requests require a JWT token in the Authorization header. Obtain a token by logging in.

POST /generis/api/auth/login

Authenticate and receive a JWT token.

Request
{ "email": "user@example.com", "password": "yourpassword" }
Response
{ "token": "eyJhbGci...", "user": { "id": "u_123", "email": "user@example.com" } }
Usage (curl)
curl -H "Authorization: Bearer <token>" \ https://generistech.com/api/agents

Base URL: https://generistech.com/api

🤖 Agents API

GET/agents

List all agents for the authenticated user.

Response
[{ "id": "agent_abc", "name": "Support Bot", "type": "customer_support", "model": "claude-sonnet", "status": "active" }]
POST/agents

Create a new agent.

Request
{ "name": "My Agent", "type": "customer_support", "system_prompt": "You are a helpful assistant...", "model": "claude-sonnet", "temperature": 0.7 }
GET/agents/:id

Get a specific agent by ID.

PUT/agents/:id

Update agent settings. Same body as POST.

DELETE/agents/:id

Permanently delete an agent and all its data.

💬 Chat API

POST/agents/:id/chat

Send a message to an agent and receive a response.

Request
{ "message": "Hello, how can you help me?", "conversation_id": "conv_xyz" // optional }
Response
{ "response": "I can help you with...", "conversation_id": "conv_xyz", "tokens_used": 142 }

🗨️ Conversations API

GET/agents/:id/conversations

List all conversations for an agent.

Response
[{ "id": "conv_xyz", "visitor_name": "John Doe", "message_count": 12, "created_at": "2025-01-15T10:30:00Z" }]

📚 Knowledge Base API

GET/knowledge-bases

List all knowledge bases (filter by agent_id in query params).

POST/knowledge-bases

Create a new knowledge base.

Request
{ "name": "Product FAQ", "agent_id": "agent_abc" }
GET/knowledge-bases/:id/documents

List documents in a knowledge base.

Response
[{ "id": "doc_1", "title": "Return Policy", "content": "Our return policy..." }]
POST/knowledge-bases/:id/documents

Add a document to a knowledge base.

Request
{ "title": "Return Policy", "content": "Full document text here..." }
DELETE/knowledge-bases/:id/documents/:docId

Delete a specific document.

🔧 Tools API

GET/agents/:id/tools

Get tools configuration for an agent.

Response
{ "tools": [ {"name": "web_search", "enabled": true}, {"name": "calculator", "enabled": false} ] }
PUT/agents/:id/tools

Update enabled tools for an agent.

Request
{ "tools": ["web_search", "calculator", "current_datetime"] }

Available tools: web_search, calculator, current_datetime, http_request, memory_save, memory_recall, knowledge_base_search, send_email, format_response, set_variable

Anthropic server-side tools (Claude agents only): anthropic_web_search (web search with cited sources, $10 per 1,000 searches; replaces the built-in web_search on Claude when both are enabled), anthropic_web_fetch (fetches a page or PDF whose URL appeared in the conversation; token cost only). Optional limits live in the agent config: web_search.max_uses, web_search.allowed_domains or blocked_domains, web_search.user_location, web_fetch.max_uses, web_fetch.max_content_tokens, web_fetch.allowed_domains.

PUT/agents/:id/advisor

Give the agent an advisor: a stronger Claude model it consults mid-reply for a plan or course correction (Anthropic advisor tool). Off by default. Consultations are billed at the advisor model's rates and appear in the conversation tool log as advisor.

Request
{ "enabled": true, "model": "claude-opus-5-5", // must be at least as capable as the agent model "max_uses": 2, // per reply, 1-5 "max_tokens": 2048 // advisor output cap per consultation, 1024-8192 }

Valid pairs: Sonnet 5.5 or Haiku 5.5 agents can use Opus 5.5, Fable 5.1 or Sonnet 5.5; Opus 5.5 agents can use Fable 5.1 or Opus 5.5; Fable 5.1 agents only Fable 5.1. An invalid model is replaced by the default for the agent model. Also applies to agents running on Claude Managed Agents.

📅 Schedules API

GET/agents/:id/schedules

List all schedules for an agent.

Response
[{ "id": "sch_1", "label": "Morning Report", "cron_expr": "0 9 * * *", "prompt": "Generate today's summary", "active": true, "last_run_at": "2025-01-15T09:00:00Z", "last_run_status": "success" }]
POST/agents/:id/schedules

Create a new schedule.

Request
{ "label": "Morning Report", "cron_expr": "0 9 * * *", "prompt": "Generate today's summary" }
PATCH/schedules/:id

Update schedule (e.g. toggle active).

Request
{ "active": false }
DELETE/schedules/:id

Delete a schedule permanently.

POST/schedules/:id/run

Manually trigger a schedule immediately.

🪟 Widget Embed

Embed a chat widget on any website with a single script tag.

Basic Embed
<script> window.GenerisConfig = { agentId: 'your-agent-id', position: 'bottom-right', theme: 'dark', primaryColor: '#6c5ce7' }; </script> <script src="https://generistech.com/widget/widget.js"></script>
Config Options
{ "agentId": "string (required)", "position": "'bottom-right' | 'bottom-left'", "theme": "'dark' | 'light'", "primaryColor": "'#6c5ce7' (hex)", "welcomeMessage": "string", "hideOnMobile": false }

⌨️ CLI

The generis-cli manages agents from the terminal. Source: Documents/GenerisAI/generis-cli. Install locally with npm link. npm publish is pending.

Installation
cd ~/Documents/GenerisAI/generis-cli && npm link
Planned commands
generis login # authenticate generis agents list # list all agents generis agents create --name "Bot" # create agent generis chat <agentId> # interactive chat generis agents delete <agentId> # delete agent generis config set api-url <url> # set custom API generis --help # show all commands