# Xberg Pricing

Two options: self-host the open-source library for free, or apply to the invitation-only Design Partner program for the managed Xberg Enterprise cloud platform. Xberg Pro — a self-hosted single-container product — sits between the two for smaller extraction workloads.

## Open Source — Self-hosted

- **Price**: Free
- **License**: MIT. Free for personal, internal, and commercial use. The only requirement is preserving the copyright notice.
- **What you get**:
  - Full Rust core
  - SDKs for 15 languages: Rust, Python, Node.js, WebAssembly, Java, Go, C#, PHP, Ruby, Elixir, Dart, Kotlin, Swift, Zig, C (FFI)
  - Docker images
  - CLI
  - Full control over your infrastructure and data, including OCR backend choice (Tesseract, PaddleOCR, Candle, or VLM) and arbitrary embedding providers
- **Source**: <https://github.com/xberg-io/xberg>

## Xberg Pro — Self-hosted, single container

- **Price**: Contact <contact@xberg.io> for availability and terms.
- **What you get**: the same extraction engine as a single-container deployment for smaller workloads — in-process job dispatch (no NATS), local disk cache, Prometheus metrics. No distributed-systems overhead.

## Design Partner program — Xberg Enterprise

- **Price**: By application. Disclosed during the application conversation, then locked for 24 months.
- **Access**: Invitation-only, curated cohort. Free-tier access may be available at our discretion for non-profits, research, and early-stage organizations in exchange for attribution.
- **What you get**:
  - Managed, multi-tenant cloud platform: REST API, dashboards, signed webhooks, usage analytics and quota dashboard
  - 101 file formats (115 file extensions) and 371 code languages
  - Images and scanned PDFs — Tesseract, PaddleOCR, Candle, and VLM OCR backends
  - Layout detection and table extraction
  - Chunking and embeddings via `extraction_config`
  - Direct roadmap input and early access to new pipeline stages
- **Apply**: <https://xberg.io/apply> or <contact@xberg.io>

## Data handling

Documents are processed in memory and deleted immediately after extraction. No storage, no indexing. Your data is never used to train models.

## See also

- [Capabilities](https://xberg.io/llms/capabilities.md)
- [Extraction API](https://xberg.io/llms/api.md)
- [FAQ](https://xberg.io/llms/faq.md)
