Welcome to Flexor

Flexor is an agentic workflow platform built for alternative asset managers. It lets your operations team automate recurring data tasks - report generation, reconciliation alerts, data quality checks, LP notification workflows - without writing production code.

The mental model

Flexor sits one layer above your data infrastructure. It does not store your data. It reads from and writes to systems that already exist in your environment: your data platform, your file store, your communication tools. What Flexor adds is the coordination layer - the logic that says “when X happens in the data, do Y, then notify Z.”

There are four primitives:

• Vaults - structured directories that hold your configuration, rules, and knowledge. Think of them as the brain of the system: everything Flexor knows about how your firm works lives in a vault.
• Skills - reusable units of capability. A skill knows how to do one thing: pull a data feed, format a report, run a check, send a message. Skills are composable.
• Agents - the orchestrators. An agent watches for a trigger (a schedule, an event, a data condition), then runs a sequence of skills to produce an output.
• Knowledge base - the memory layer. Agents consult the knowledge base to answer questions, resolve ambiguity, and improve over time.

What your first deployment looks like

Most teams start with one agent automating one workflow. A common first deployment at a hedge fund:

1. An agent monitors data freshness in the PlexiFact catalog. If a feed is more than 30 minutes late, it sends a structured alert to the operations Slack channel with the feed name, expected time, and the last successful ingestion timestamp.
2. The same agent logs the event to an audit trail in the vault.

That agent requires no code. It requires a vault configuration, a data freshness skill (available in the standard library), and a notification skill. Setup time: approximately 90 minutes for an operator who has completed the pre-work.

How this documentation is organized

Section	What it covers
Get Started	Installation, quickstart, and core concepts
Vaults	The four-tier model, configuration, and best practices
Skills	Anatomy, composition, and design patterns
Agents	Overview, custom builds, and multi-agent coordination
Operations	Deployment, governance, and troubleshooting
Reference	CLI commands and every configuration key

If you are new to Flexor, start with Concepts after this page, then move to the Quickstart.

Note: This documentation describes the Flexor specification. Some CLI commands and configuration keys are marked preview - they are specified and designed but implementation is in progress. Preview items are flagged explicitly.