DreamFactory review

DreamFactory is a self-hosted server and API gateway, not a hosted API service, client library or command-line tool. It sits between applications or local and on-prem LLMs and the data sources they need to access. The platform generates standardized REST APIs over databases, file stores, SaaS and legacy systems, and also provides MCP access for AI agents.

That layer centralizes authentication, role-based access, request controls and audit visibility instead of giving each application or model direct database credentials. The vendor says credentials stay within DreamFactory and that the platform maps to existing security infrastructure and preserves roles and entitlements. Deployment options include on-prem, private cloud, edge and hybrid environments.

For a team, the architectural tradeoff is clear: DreamFactory adds a server to operate, but can replace hand-built connectors and place a consistent policy boundary around varied backends. It is aimed at governed data access and API generation, rather than general-purpose API hosting.

The primary integration surface is generated REST APIs: DreamFactory creates a consistent API layer over connected data sources. For AI workflows, the vendor also describes MCP access, with MCP tools backed by DreamFactory APIs. Its homepage lists more than 30 connectors, including PostgreSQL, MySQL, SQL Server, Oracle, DynamoDB, MongoDB, S3, Azure Blob, SFTP, Salesforce, REST and SOAP services.

The homepage links to an API Generation guide and API Management concepts, and describes automatic API creation and testing. Its stated gateway controls include routing, authentication, rate limits and observability. The vendor homepage does not name programming-language SDKs or a CLI. Teams planning to build around a particular language should confirm SDK support in the vendor's docs and check how generated API schemas and authentication fit their clients.

This makes the integration model API-first: connect a backend, configure access, then call its generated endpoints. Confirm connector-specific behavior, write support and MCP requirements before choosing a production architecture.

The Open Source plan is free and includes a PostgreSQL connector. The pricing page lists DF Linux Lite at $1,500/month, billed annually, with one database connector and unlimited roles, keys, API creation and requests. DF Linux Professional is $4,000/month, billed annually; it adds unlimited connectors and includes all connectors except Oracle, Snowflake and DB2.

Docker/Kubernetes plans start at $6,000/month, billed annually, at 4 vCPU, with each additional vCPU listed at $1,500. They include all connectors and API scripting, authentication, rate limiting, network APIs, logging and governance. Custom-priced enterprise and federal plans are also listed. The page gives no separate month-to-month prices.

The listed plans specify unlimited requests, so the published pricing does not describe per-request overages. Compute can still raise the bill on Docker/Kubernetes deployments as vCPU needs grow. Support is an add-on: Basic is listed at $500/month billed annually; Pro and Enterprise each show two prices on the page, so confirm the applicable rate and terms with the vendor before budgeting.

DreamFactory offers a free plan labeled Open Source and paid self-hosted deployments, including Linux, Docker and Kubernetes options. The vendor's homepage describes on-prem, private cloud, edge and hybrid deployment, so the product is not limited to a vendor-hosted service.

The vendor does not publish the software license terms on the homepage. Confirm which components and connectors are covered by the license, and whether the free plan permits your intended production use, before adopting it. A plan labeled Open Source does not by itself establish the license or source availability for every edition.

DreamFactory creates REST interfaces over existing systems, but the vendor does not describe a data-export or configuration-export process on its homepage. Your source data remains in its connected systems, but moving the API layer means recreating endpoint, policy and connector configuration elsewhere. Ask about export formats and migration support before building extensive service-specific configuration.

DreamFactory lists rate limits, throttling, quotas, logging, observability, request validation and threat protection as gateway controls. Its identity and role features are intended to keep existing entitlements in force, and the vendor positions the platform for sensitive enterprise data. Teams should validate connector behavior, authorization rules, audit coverage and capacity under their own workloads before putting production traffic behind it.

The pricing page lists unlimited requests for several paid plans, but does not publish request-rate ceilings or service-level commitments. Docker/Kubernetes pricing starts at 4 vCPU and charges $1,500 per additional vCPU, so deployment sizing affects cost. Ask the vendor for supported scaling patterns, operational requirements and reliability commitments for your chosen plan.

A Reddit user reports that renaming a service broke an API; another commenter says renaming can work after flushing the affected service cache. Treat cache behavior and configuration changes as release checks, and test rename or deployment workflows in a non-production environment before relying on them.