HiveLang v5
The language of agents
A declarative language for orchestrating autonomous AI agents. It turns prompts, tools, memory, schedules, and recovery behavior into readable runtime code.
Documentation
guide
Syntax Basics
Variables, bot definitions, and core logic.
guide
Parallel Execution
Concurrent tool calls and error boundaries.
guide
Standard Library
Built-in tools, AI, and HTTP capabilities.
guide
Memory
Store preferences, seen items, user style, and durable context.
guide
Schedules
Run agents on cron without making users manually trigger them.
guide
Retries & recovery
Handle flaky APIs, failed tools, and safe fallback paths.
guide
Production patterns
Real agent patterns for support, research, content, and ops.
Why HiveLang?
Declarative Logic
Instead of fragile prompts, define bot behavior with structured statements like on user.message and if/else.
Native Multi-Agent Support
Define multiple specialized agents in one swarm and let the runtime handle orchestration.
Integrated Memory
Built-in db.get and db.set commands for persistent agent state without extra boilerplate.
Concurrent Tools
The parallel keyword makes multi-tool execution blazingly fast.
Example
bot AutoResearcher {
instructions {
Finds news on a topic and summarizes it.
}
capabilities {
web.search
news.search
ai.generate
}
on user.message {
parallel {
call web.search with { query: input } as search_data
call news.search with { query: input } as news_data
}
call ai.generate with {
prompt: "Summarize " + search_data + " and " + news_data + " for the user."
} as final_summary
respond with final_summary
}
}