Whatever message this page gives is out now! Go check it out!
ColdFusion 2025 update 8 adds first-class AI support directly into the language. You access AI providers through built-in CFML functions that integrate with your existing components, built-in functions, and application logic, not through generic HTTP calls. A unified provider framework lets you write one set of code that works across multiple AI providers: OpenAI, Anthropic, Mistral, Google Gemini, and Ollama. To switch providers, you update your configuration. No code refactoring is required. View Introduction to models in ColdFusion for more information. |
Model Context Protocol (MCP) is an open standard for connecting AI models to external systems. It provides a controlled interface layer between an AI agent and your application, giving agents the ability to discover and use tools without unrestricted access to your codebase. ColdFusion 2025 update 8 supports MCP in both directions as client and server. View MCPs in ColdFusion for more information. |
ColdFusion 2025 update 8 adds a native RAG pipeline that grounds AI responses in your own content. It supports five vector stores (InMemory, Pinecone, Milvus, Qdrant, Chroma) and offers three approaches: simpleRAG() for quick setup, agent() with INGESTION for full pipeline control in production, and documentService() for scheduled or shared ingestion jobs.
|
ColdFusion 2025 update 8 adds five security improvements: native passkey support (WebAuthn/FIDO2) for passwordless sign-in, an Argon2-default password hashing API, an expanded Security Analyzer that now detects unscoped variable risks and exports PDF/CSV reports, OEM library upgrades addressing underlying vulnerabilities, and AI guardrails in agent() to filter LLM inputs and outputs via CFC-based validators.
|
The ColdFusion Builder VS Code extension now supports Linux and Docker (as a first-class server target with full RDS/debugger support), has a rebuilt context-aware CFML grammar that no longer breaks syntax highlighting on mixed-language files, adds code completions for AI, Cloud, and MongoDB services with deprecation warnings, and ships an improved Security Analyzer with new rules, better performance on large codebases, and PDF/CSV export. View What's new and changed in ColdFusion Builder VS Code Extension Update 5 for more information. |
PMT now includes an AI Services dashboard (enabled via Settings > Monitoring > Advanced) with seven tabs covering every layer of the AI stack; LLMs, Agents, RAG, MCP Clients, MCP Servers, Vector Stores, and a Trace Viewer that shows a full flamegraph of each AI request to pinpoint pipeline bottlenecks. View AI services monitoring in PMT for more information. |
ColdFusion 2025 update 8 adds native Sets with set theory operations, CompletableFuture-based async with asyncAllOf(), asyncAnyOf(), and timeout/combining methods, automatic CFML-to-Java functional interface adaptation, member functions on literals, and eight new or enhanced built-in functions. View the following for more information:
|
| If you have a question to ask or an idea to share, come and participate in the Adobe ColdFusion Community . We would love to hear from you and address any queries you may have. |