Alquimia Platform

Why enterprises choose Alquimia Platform

Most AI agents fail in production — not because the models are weak, but because the platform around them lacks the controls that enterprises require: auditability, governance, security, and reliable integration with existing systems.

Alquimia Platform is the agent management layer that closes this gap. It wraps LLMs and autonomous agents in explicit contracts, deterministic execution paths, and enterprise-grade controls — so your AI investments deliver measurable outcomes, not just demos.

Governed by design

Every agent is defined by a versioned, validated configuration. Nothing is inferred at runtime. Hold, release, dry-run, and audit agent behavior before it reaches production.

Auditable execution

Every agent action is a typed event in an append-only worklog. Full execution traces for compliance, debugging, and replay — no black boxes.

Enterprise security

Keycloak OIDC, JWT, and API-token auth. HashiCorp Vault for secrets. OCI artifact distribution. Shield enforcement for prompt injection and unsafe content. Zero Trust workload identity optional.

Multi-model, multi-channel

Connect OpenAI, Anthropic, Ollama, vLLM, or any OpenAI-compatible endpoint. Receive messages from WhatsApp, Slack, Email, and web — normalized through the same execution pipeline.

Built for two audiences

Alquimia Platform is both an application platform and an infrastructure platform. Application teams use it to deliver agent-powered experiences, while platform and SRE teams operate it with the controls they expect.

Application teams

Build agents that reason over knowledge, call enterprise APIs, and converse across WhatsApp, Slack, Email, and web. Focus on behavior and business logic, not plumbing.

Platform & SRE teams

Deploy a horizontally scalable, Kubernetes-native service with Vault-backed secrets, workload identity, signed images, and correlated telemetry.

Built as an open platform

Alquimia Platform ships as three components that work together or independently:

• alquimia-core — The agent execution engine. Evaluation strategies, memory management, tool integration, agent registry, knowledge sources, and shield enforcement. See the Core SDK overview.
• alquimia-runtime — The HTTP service layer. Async inference, SSE streaming, session state, authentication, and registry management. See the Runtime API overview.
• alquimia-studio — The web-based studio for designing, testing, and managing agents, registries, and knowledge sources through a visual interface. See the Studio overview.

In a typical deployment, the runtime runs as a Kubernetes workload and uses the core engine internally. Applications and channels interact with it over HTTP. For embedded or custom Python applications, the core SDK can be used directly. Studio connects to the runtime to provide a visual management layer.

Explore the platform

What is Alquimia Platform? The production-AI gap, core principles, and how the platform fits into your stack.

Architecture Event-driven execution, master/worker topology, and request flow.

Use cases How enterprises deploy Alquimia Platform for support, IT ops, documents, and multi-agent workflows.

How-To Guides Step-by-step recipes for agents, channels, tools, approvals, secrets, scaling, and audit.

Security & Zero Trust Authentication, Vault policies, workload hardening, and audit.

Troubleshooting Symptom-based guides for wrong answers, stuck tasks, credentials, channels, performance, and alerts.

Platform components

Component	Documentation	Description
alquimia-core	Core SDK overview	Python SDK — agent execution engine, memory, tools, registry, shields
alquimia-runtime	Runtime API overview	FastAPI service — HTTP API, async inference, streaming, auth, Vault
alquimia-studio	Studio overview	Web UI — design, test, publish, and operate agents, registries, and knowledge sources
alquimia-trust-lens (optional)	TrustLens overview	Conversation evaluation service — Gaussia benchmarks and Agent Humanity metrics for Studio