Skip to main content
An agent is a configurable AI assistant. It combines a language model with instructions, optional tools, memory, and knowledge — packaged as a reusable unit that the Alquimia Runtime can execute.

​
What makes up an agent

ComponentWhat it doesRequired
Language modelThe LLM that generates responsesYes
System promptIdentity, personality, rules, and purposeYes (auto-generated or custom)
MCP toolsExternal services the agent can callNo
Agent-to-AgentOther agents the agent can delegate toNo
MemoryConversation history (short-term) or persistent context (long-term)No
Knowledge baseDocument collection for RAG-based answersNo
Integration channelsWhatsApp, Email — external communication interfacesNo
Pre-processorSentinels and empathy rules that run before the agentNo

​
How agents run

When a user sends a message: 
User message
    ↓
Pre-processor (sentinels, empathy rules) — optional
    ↓
Memory retrieval — inject relevant history
    ↓
Knowledge retrieval — inject relevant document passages
    ↓
LLM call — prompt + context + tool schemas
    ↓
Tool execution (if needed) — ReAct loop
    ↓
Response returned to user
The Runtime handles all of this automatically based on the agent’s configuration. You don’t write any of this logic — you configure it in Studio.

​
Evaluation strategies

The evaluation strategy controls how the agent thinks. This is one of the most important settings for an agent.
StrategyDescriptionBest for
offSingle LLM call, no reasoning loopSimple Q&A, content generation
one-shootOne call with tool schemas injectedStructured output, simple tool use
toolsLLM calls tools in a single roundDeterministic, tool-first workflows
reactReAct loop: reason → act → observe → repeatMulti-step tool use, complex tasks
long-taskExtended planning for complex, multi-step tasksResearch agents, autonomous workflows
When you add MCP servers or other agents to your agent, Studio automatically switches the evaluation strategy to react. This ensures the agent can actually use its tools. You can override this manually in Dev Mode.

​
The ReAct loop explained

react is the most powerful strategy. The agent follows this cycle:
  1. Reason — “I need to look up the user’s account to answer this”
  2. Act — calls the get_account tool from an MCP server
  3. Observe — receives the tool result
  4. Reason again — “Now I have the account info, I can answer”
  5. Respond — generates the final answer
This continues until the agent has enough information to respond or hits its reasoning limit.

​
Next steps

Creating an Agent

Step-by-step walkthrough of the agent creation wizard.

Dev Mode

Take full control of the system prompt and evaluation strategy.