James Miller
An open-source clinical backend for building AI health apps with FHIR R4 support and real-time sync capabilities.
Questions about bonfireDB
What is bonfireDB?
bonfireDB is an open-source clinical backend designed for building AI health applications, using TypeScript and Postgres with a canonical FHIR R4 data model. It supports both patient-facing and clinician-facing surfaces on the same backend, featuring real-time bidirectional syncing and no indexing delay on data commits. The platform includes FHIR R4 export, audit logging, provenance tracking, and attribute-based access control, along with a permission-scoped evidence packet for AI agents. The project is in early access with a planned Apache-2.0 license.
What are the pricing options and plans for bonfireDB?
The Open Source tier is planned as free under Apache-2.0 and runs entirely in your own AWS, while the Managed Cloud tier is in early access and carries a Business Associate Agreement (BAA). The Enterprise/BYOC tier is also in early access and requires talking to the team for details. Paid-tier pricing for the managed and enterprise levels is not yet finalized.
Who is the ideal customer for bonfireDB?
bonfireDB is designed for teams building AI health apps who need a clinical context layer without managing a separate FHIR server. It suits those comfortable operating their own AWS infrastructure or those who want a partner to carry the BAA and handle backups via the Managed Cloud tier. The core is built for patient-scoped context compilation, consent, and bounded traversal.
How does the bonfireDB workflow operate from setup to execution?
The process begins by scaffolding a FHIR-safe backend with a typed SDK using a single command, then deploying it into your own AWS environment. You can then use app-native primitives like clinical.notes.create() instead of raw FHIR REST, and utilize operational read models and reactive queries. The system automatically handles FHIR R4 export, audit, provenance, and ABAC underneath the application layer.
What specific technical features does bonfireDB offer for AI health applications?
bonfireDB features conformance-tested projections, cited search, and governed agent proposals in its core today. It also provides real-time bidirectional syncing where a clinician's check-in updates the patient app instantly, and a patient logging a tic appears in the clinician's timeline immediately. The system includes a permission-scoped evidence packet that reduces raw FHIR token dumps to a scoped, cited packet for agents.
How does bonfireDB handle data synchronization between patient and clinician interfaces?
The platform distinguishes itself by offering a single backend that powers both a patient app and a clinician portal, ensuring they are always in sync by default. Unlike typical FHIR backends that wait for search indexes, bonfireDB ensures fresh data on every commit, so a patient's tic is visible to a clinician instantly. This real-time capability works in both directions, allowing a doctor writing to their own patient to enforce clinical authorization.
What is the current development status and release timeline for bonfireDB?
bonfireDB is currently in early access, with the core planned as open source and pre-release today. The project is described as early-stage, and the broader contract for patient-scoped context compilation, consent, and bounded traversal is still being built and tested in the open. The website notes that pages reflect product design and positioning, not all shipped features.
Which technologies and cloud environments does bonfireDB integrate with?
The platform integrates with Postgres and pgvector, and allows deployment into your own AWS environment. It is designed to run without a Business Associate relationship for the open-source core, though the Managed Cloud tier includes a BAA backed by their AWS BAA. The system also supports community support through Discord and GitHub.
What are the key differences between bonfireDB's approach and standard FHIR servers like HAPI or Medplum?
bonfireDB aims to replace raw FHIR REST calls with app-native primitives like clinical.notes.create() for a cleaner experience. It provides typed reads on top of canonical FHIR R4/core profiles stored in JSONB, and offers clean Bundle export functionality. The roadmap includes agent context tools where typed functions become SDK and MCP tools with a complete HTTP surface.
How does bonfireDB optimize data usage for AI agents compared to raw FHIR exports?
The system includes a permission-scoped evidence packet that generates approximately 4,200 tokens compared to a raw FHIR dump of around 3,000,000 tokens. This scoped packet is cited and query-aware, allowing agents to read evidence without overwhelming context limits. The packet includes Observation and QuestionnaireResponse citations for proposed messages awaiting approval.
What security and compliance measures are included in the bonfireDB Managed Cloud tier?
The platform includes automated backups and point-in-time recovery for the Managed Cloud tier, along with audit-log retention and monitoring and alerting. It also offers team management capabilities and is backed by an AWS BAA for the managed service. The core is planned as open source with community support, while the managed tier handles operational tasks.
What is the official website of bonfireDB?
The official website of bonfireDB is bonfiredb.dev.
How is bonfireDB rated on Trustburn?
On Trustburn, bonfireDB has a rating of 3.8 out of 5 based on 5 reviews.
AI-generated from reviews on this page. Verify anything important with the company directly.
Sarah Jenkins
The quickstart workflow looks incredibly streamlined. Scaffolding a backend with typed SDKs and deploying it to AWS seems like a massive time-saver for teams building clinical apps.
David Chen
The open-source clinical backend appears well-architected for modern AI health applications. The emphasis on canonical FHIR R4 with immutable history provides necessary data integrity. Key differentiators like the fresh-on-commit architecture and real-time bidirectional sync address common latency issues in traditional systems. The provision of typed SDKs and conformance-tested projections in Postgres simplifies development compared to managing raw FHIR servers. Although some advanced features like patient-scoped context compilation are still in progress, the current core tools for cited search and agent context offer a practical starting point. The inclusion of automatic audit, provenance, and ABAC underneath the surface reduces overhead for developers building compliant applications.
Elena Vasquez
From the information available, the real-time synchronization between patient and clinician surfaces is quite impressive. The claim that a tic logged by a patient appears instantly in the clinician's timeline, without the typical indexing delays found in standard FHIR backends, suggests a robust underlying architecture. This immediate visibility, combined with the governed agent proposals and cited search capabilities, offers a significant convenience for clinical workflows. The ability to deploy a FHIR-safe backend with a single command into AWS also reduces initial setup friction. While some features like semantic embeddings are still on the roadmap, the current core implementation of canonical FHIR R4 and automatic audit trails provides a strong foundation. For someone considering this company, the focus on clinical context layers rather than just raw data storage seems to address specific pain points in AI health app development effectively.
Marcus Thorne
The technical capabilities look solid, but from the information available, the service offering seems to lack clear pricing transparency for developers considering integration costs.
News
No news yet. This is where the business shares its own updates and press.
Photos
No photos yet. Photos shared by reviewers will show up here.
Widgets
Get the widget for your site →About bonfireDB
bonfireDB is an open-source clinical backend designed for building AI health applications. It uses TypeScript and Postgres with a canonical FHIR R4 data model underneath, and includes features such as conformance-tested projections, cited search, and governed agent proposals. The platform supports both patient-facing and clinician-facing surfaces on the same backend, with real-time bidirectional syncing and no indexing delay on data commits. It also provides FHIR R4 export, audit logging, provenance tracking, and attribute-based access control, along with a permission-scoped evidence packet for AI agents. The project is in early access, with a planned Apache-2.0 license, and is seeking design partners.
- Website
- bonfiredb.dev