Skip to content
github-actions[bot] edited this page May 21, 2026 · 2 revisions

Frequently Asked Questions

Common questions about Promptly, answered for both business and technical audiences.


General

What is Promptly?

Promptly is an open-source platform for managing, securing, and governing the AI prompts that power your organization's AI agents and applications. Think of it as version control + approval workflows + security scanning — purpose-built for AI prompts.

Is Promptly free?

Yes. Promptly is open source under the Apache License 2.0. You can use, modify, and deploy it without licensing fees.

Do I need to be a developer to use Promptly?

No. Promptly is designed so that product managers, compliance officers, and business users can author prompts, review changes, and monitor AI behavior through a web interface. Developers set up the initial deployment and integrations.

What AI models does Promptly support?

Promptly supports multiple LLM providers out of the box: OpenAI (GPT-4, etc.), Google Gemini, Anthropic (Claude), and Ollama (for local/self-hosted models). You can switch providers via configuration — no code changes required.


Security & Compliance

How does the vulnerability scanner work?

When a prompt is created or edited, Promptly's AI-powered scanner analyzes it for common risks: prompt injection vulnerabilities, PII/PHI data exposure, missing safety guardrails, and toxicity concerns. Critical findings block the prompt from being submitted for review.

Is Promptly compliant with SOC 2 / HIPAA?

Promptly provides the tooling that supports compliance — immutable audit logs, role-based access control, approval workflows, and security scanning. Whether your deployment meets a specific standard depends on your overall infrastructure and organizational controls. Promptly gives you the evidence trail auditors look for.

Can I self-host Promptly?

Yes. Promptly is designed to be fully self-hosted. Run it on your own servers, your own cloud, or behind your firewall. Your prompt data never leaves your network unless you choose to use a cloud LLM provider for scanning/improvement features.

Does Promptly send my prompts to external AI services?

Only when you explicitly use the AI-powered features (vulnerability scanning and quality improvement). These features call the configured LLM provider. If data sovereignty is a concern, configure Promptly to use Ollama with a locally-hosted model — all processing stays on your infrastructure.


Deployment & Integration

How do my AI agents get prompts from Promptly?

Your agents call the Runtime Delivery API: GET /api/v1/deliver?appId=X&usecase=Y&agent=Z. Promptly returns the latest approved and deployed prompt version. SDKs are available for TypeScript, Python, and Java to simplify integration.

Can I use Promptly in my CI/CD pipeline?

Yes. Promptly's Export/Import APIs are designed for CI/CD integration. Export prompts from your development instance as a JSON bundle, then import them into staging or production instances. Works with GitHub Actions, GitLab CI, Jenkins, and any pipeline that can make HTTP calls.

What databases does Promptly support?

Promptly supports MongoDB (default, with Atlas Vector Search), PostgreSQL (with pgvector), and H2 (embedded, for local development and demos). Switch between them using Maven profiles and Spring configuration.

How do I update prompts without downtime?

That's one of Promptly's core features. Author and approve a new prompt version through the web UI. Once deployed, your AI agents automatically receive the updated prompt on their next Runtime API call — no application redeployment needed.


Team & Workflow

How does the approval workflow work?

Prompts follow a structured lifecycle: Draft → In Review → Approved → Deployed. An author creates or edits a prompt (Draft), submits it for review, and a designated Reviewer approves or rejects it. Only approved prompts can be deployed to production.

Can I skip the approval workflow?

Project Admins have the ability to bypass the workflow for emergency situations. However, all actions — including bypasses — are recorded in the immutable audit trail for accountability.

How many projects can I create?

There is no limit. Create as many projects as you need to organize prompts by product, team, environment, or any structure that works for your organization.

Who gets notified when a prompt needs review?

Promptly uses real-time notifications (Server-Sent Events) to alert Reviewers and Admins when a prompt is submitted for review. Notifications appear in the app's notification bell.


Technical

What tech stack does Promptly use?

  • Frontend: Angular 21 · TypeScript · Angular Material 3 · NgRx
  • Backend: Java 21 · Spring Boot 4.0 · Spring WebFlux (reactive)
  • AI: Spring AI (multi-provider support)
  • Build: Nx monorepo · Maven · pnpm
  • Containers: Docker & Docker Compose

Is Promptly a microservices architecture?

No. Promptly uses a modular monolith architecture powered by Spring Modulith. This provides the organizational benefits of microservices (clear module boundaries, independent domain models, event-driven communication) with the operational simplicity of a single deployable unit.

Can I contribute to Promptly?

Absolutely! Promptly welcomes contributions — bug reports, feature requests, documentation improvements, and code. See the Contributing guide for details.


Have a question not covered here? Open a Discussion on GitHub.

Clone this wiki locally