Now in Beta

Build AI agents,
visually.

Drag blocks onto a canvas. Connect LLMs to your data, tools, and APIs. Test in real time. Deploy as a chat link or REST API — all without writing orchestration code.

BLLM Agent Editor
SavedRun
Start
Entry point
Input
Ask question
LLM
GPT-4o
Switch
Route intent

The visual editor — drag blocks, connect them, and build any AI workflow.

18 building blocks, infinite possibilities

Each block is a step in your agent's workflow. Connect them to define how your agent thinks, acts, and responds.

LLM

Send a prompt to any model. Supports tool calling.

Tool

Web search, SQL queries, RAG, MCP, custom JS.

Switch

Multi-way routing based on intents or variables.

Loop

Iterate over arrays and process lists of data.

Memory

Save and load state across user sessions.

Validator

Ensure data matches regex, emails, or ranges.

Transform

Format JSON, extract fields, and manipulate text.

Sub-Agent

Chain specialized agents together for complex tasks.

Handoff

Pause the AI and seamlessly escalate to a human.

A/B Split

Randomly split traffic to test different prompts.

Code

Run sandboxed JavaScript to transform data.

HTTP Request

Call any REST endpoint and parse the response.

From idea to live agent in 4 steps

No infrastructure to manage. No code to write. Just build, test, and ship.

1

Design your flow

Open the editor. Pick a template or start blank. Drag blocks onto the canvas — Start, LLM, Tool, Output — and connect them with edges.

Templates available: Customer Support, Order Tracker, Website Analyzer, Product Recommender.

2

Configure each block

Click any block to configure it. Choose your LLM provider and model, write prompts with {{variables}}, set up conditions, or paste custom code.

Supports OpenAI, Anthropic, Google Gemini, DeepSeek, xAI Grok, Mistral, Groq, and OpenRouter.

3

Test in the playground

Hit Run. A chat panel opens on the right. Talk to your agent, watch the flow execute in real time — the active block highlights green.

Fix issues instantly. No deploy step needed to test.

4

Deploy and share

Publish your agent. Share it as a public chat link, embed it in your site, or integrate via the REST API with API key authentication.

Conversations persist. Users can bookmark and resume sessions.

Enterprise orchestration, built-in

Everything you need to build robust, scalable AI agents without writing orchestration code.

Advanced Routing

Go beyond simple true/false conditions. Use multi-way switches, loops, and A/B testing to build complex, non-linear flows.

State & Memory

Agents remember user preferences across sessions. Save, load, and transform data visually without setting up a database.

Agent Composability

Don't build one massive god-bot. Create specialized micro-agents and orchestrate them together using Sub-Agent nodes.

Human-in-the-Loop

AI can't solve everything. Seamlessly escalate to human agents when needed, pausing the bot while your team takes over.

Multi-Provider LLM

Bring your own API keys for OpenAI, Anthropic, Google Gemini, DeepSeek, xAI Grok, Mistral, Groq, or OpenRouter.

Data Guardrails

Validate user input and LLM outputs in real-time. Ensure emails, URLs, and JSON structures are correct before proceeding.

Web Browsing & Scraping

Equip agents with live web search and page scraping. Get real-time information beyond training data cutoffs.

Custom Code & MCP

Run JavaScript in sandboxed nodes. Connect to MCP servers for tool calling. Build anything the built-in blocks can't do.

REST API & Integrations

Full public API with API key auth, scoped permissions, rate limiting, and usage analytics. Integrate agents into any product.

What you can build

Real workflows running on BLLM today. Start from a template or build your own from scratch.

Support

24/7 Customer Support

Upload your docs to a knowledge base. The agent answers questions from your data, escalates to humans when unsure, and tracks every conversation in your inbox.

StartInputLLM + RAG ToolConditionOutputHandoff
Orchestration

Enterprise Router Agent

A central dispatcher that greets users, categorizes their intent, and routes them to specialized sub-agents for billing, support, or sales.

StartInputLLMSwitchSub-AgentMemory
E-commerce

Order Status Tracker

Connect your PostgreSQL database. Customers ask about their order and the agent writes a SQL query, fetches the result, and responds with live data.

StartInputLLM + SQL ToolOutput
Sales

Lead Qualification Bot

A conversational agent that asks qualifying questions, scores the lead based on conditions, and routes hot leads to your sales team via handoff.

StartInputLLMValidatorOutputHandoff
For Developers

REST API with full control

Integrate agents into your own product. Create API keys with scoped permissions, execute agents programmatically, manage conversations, and query knowledge bases — all via a versioned REST API.

  • API key auth with SHA-256 hashing
  • Scoped permissions (read, execute, write)
  • Per-key rate limiting and IP allowlists
  • Conversation persistence and history
  • Usage analytics for every request
Execute an agent
curl -X POST /api/v1/agents/:id/execute \
  -H "Authorization: Bearer bllm_pk_..." \
  -H "Content-Type: application/json" \
  -d '{
    "userInput": "What is my order status?",
    "conversationId": "conv_abc123"
  }'
Response includes events, state, and conversationId for multi-turn chats.

Works with your stack

OpenAIAnthropicGoogle GeminiDeepSeekxAI GrokMistralGroqOpenRouterPostgreSQLAstra DB

Built for production

Your data and API keys are encrypted at rest. Code runs in isolated sandboxes. Every API request is rate-limited and logged.

Encrypted secrets

LLM API keys are AES-256-CBC encrypted. Only decrypted at execution time.

Sandboxed code

Custom JavaScript blocks run in isolated-vm — no filesystem or network access.

Audit logging

Every public API call is logged with method, path, status, latency, and IP.

Ready to build your first agent?

Sign up in 10 seconds. Pick a template. Hit Run. Your agent is live.