Pirol Labs
Pirol is a small German AI lab researching human–agentic interaction and swarm intelligence. It builds tools that let people and AI agents share a desk in a predictable, observable and safe way, with a focus on privacy, human agency and sensitive data.
At a Glance
- Small and medium-sized businesses building their own operational software
- Teams and enterprises that need centralized billing, administration, role-based access, audit logs and organization-wide skills
- Developers integrating AI models through an OpenAI-compatible API
- Privacy-sensitive organizations requiring EU-hosted, local or on-premises AI infrastructure
AI Tools by Pirol Labs
(1)
Company as Code
AI Context Graph for Agents
Discussions
No discussions yet
Be the first to start a discussion about Pirol Labs
Latest News
Company as Code added charta init and npm init @pirol/charta (v0.4.0)
Company as Code added charta init to create company.yaml (v0.3.0)
Company as Code added Mermaid and DOT graph output and rendered the demo company graph (v0.2.0)
Pirol refreshed the Company as Code demo video with visible typed commands and cursor actions
Products & Services
A desktop application in which AI agents build and operate company software such as CRM, ERP and HR applications, while keeping company data on the user's machine or team server. It is available for macOS and Windows.
An open standard for describing a company's goals, roles, processes, policies and products as Markdown files in Git so people can read them, agents can follow them, and validators can check references and required structure.
The Apache-2.0 reference toolchain for Company as Code. It validates descriptions, queries and graphs the company model, plans change impact, and exposes the graph through MCP.
An OpenAI-compatible gateway for Pirol Online and Pirol Local. It provides access to Pirol AI and, using customers' own keys, OpenAI and Anthropic models.
Market Position
Pirol positions itself as a privacy-first, sovereign alternative to cloud-only AI assistants and conventional fixed SaaS: Kyte lets agents build and operate custom business software locally, while Pirol Online provides EU-hosted models and Pirol Local keeps inference on customer hardware. Company as Code differentiates the platform with a Git-native, human-readable and verifiable operating model for AI agents rather than a workflow-only UI.
Founding Story
Pirol says its motivation is that intelligence is becoming abundant and affordable, and it wants AI systems to serve people, improve lives and shorten the innovation cycle from years to days. Its practical starting point is the desktop: building agents people can trust with sensitive work, including on-premises and local deployments.
Business Model
Revenue Model
Subscription plans for the Pirol/Kyte platform, with included Pirol AI credits or token allowances; paid plans add model access, web search, Kyte Extend, transcription and higher limits. The API also supports usage through credits, while local deployment is offered for customer-owned hardware.
Pricing Tiers
About 10 million Pirol AI tokens per month; includes Kyte desktop app, Pirol Local, code and data generation, memory, files/code execution and MCP connectors.
7-day free trial; €19.95 (+5%) monthly credits; frontier models, web search, Kyte Extend and audio transcriptions.
€99 (+10%) monthly credits, higher output limits and priority access during high traffic.
Online database with backups on Pro and Scale plans.
Target Markets
- Small and medium-sized businesses building their own operational software
- Teams and enterprises that need centralized billing, administration, role-based access, audit logs and organization-wide skills
- Developers integrating AI models through an OpenAI-compatible API
- Privacy-sensitive organizations requiring EU-hosted, local or on-premises AI infrastructure
- Building and running internal CRM, ERP, HR and other custom business applications
- Automating recurring company processes such as invoicing, customer follow-up and planning
- Agentic work on sensitive or private company data where data should remain on-premises or in the EU
- Documenting company operating models so human and AI coworkers follow the same goals, roles, processes and policies
- Developing applications against a common AI gateway using Pirol, OpenAI or Anthropic models