OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Frameworks

Pragma

Pragma is a TypeScript framework and runtime for defining portable AI agent teams that combine models, tools, shared context or memory, skills, permissions, and workflows. Teams can run through Pragma Desktop or CLI, connect with Codex and Claude Code, or integrate into another AI system through an SDK.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Pragma — OpenViST project profile in Agent Frameworks.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Pragma when an AI workflow needs more structure than a single prompt and should remain portable across execution environments, models, or supported harnesses.

Who it is for

  • Developers building multi-agent applications.
  • Teams packaging repeatable AI workflows across models or harnesses.
  • Builders who need shared context, memory, skills, tools, or human review checkpoints in agent workflows.

Documented uses

  • Define reusable agent teams with experts, tools, skills, flows, and review checkpoints.
  • Coordinate multi-step work across models and supported agent harnesses.
  • Run the same team through desktop or command-line interfaces.
  • Package shared context and memory for reuse across missions.
  • Embed an agent team in a custom AI system through the SDK.
Before you choose

Requirements & limits

Pragma is public open-source software under the AGPL-3.0 license and is identified as preview software. The documented desktop release provides macOS Apple Silicon and Intel builds. Source builds require Node.js 22 or later and pnpm 10.12.1.

  • Choose it when agent teams, rather than individual prompts, are the unit you need to define and reuse.
  • Its TypeScript implementation and YAML DSL matter for developers selecting a buildable framework.
  • Desktop and CLI execution are documented separately from SDK embedding, so the intended integration route affects setup.
  • Codex and Claude Code support is described, but the supplied documentation does not specify detailed integration steps or feature limits.
  • Memory and context storage use extensible Host adapters, with several repository-provided implementations.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Build Pragma Desktop from source

Prepare and start a development build of the Pragma desktop application.

Before you start

  • A cloned copy of the public Pragma repository.
  • Node.js 22 or later.
  • pnpm 10.12.1.

Steps

  1. Clone the repository and change into its directory.

  2. Install dependencies with the repository's frozen lockfile.

  3. Start the desktop development build with the documented pnpm filter command.

Expected outcome

A development build of Pragma Desktop is started from the source repository.

Guide limitations

  • This documents setup and starting the development build, not a production packaging process.
  • Development builds use ~/.pragma-development/ by default.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
3 October 2026
Original website title
pqpo/pragma · Official homepage title · retained in its source language
How we review sourcesSuggest a correction