OpenCode v1.18.9 restores MCP 1.x compatibility

// 1h agoPRODUCT UPDATE

OpenCode v1.18.9 restores MCP 1.x compatibility

ANNOUNCEMENT PRODUCT GITHUB PRODUCT HUNT X

OpenCode released version 1.18.9, bringing back compatibility with Model Context Protocol (MCP) 1.x. The release improves app navigation and project reordering stability while introducing persistent collapsible provider groups within model settings.

// ANALYSIS

Restoring MCP 1.x compatibility is an essential fix for developers relying on standardized tool integration and protocol interoperability across coding agent environments.

• Restores full compatibility with MCP 1.x servers and extensions.

• Enhances stability for project reordering and app navigation workflows.

• Adds persistent collapsible provider groups in model settings configured per server.

// TAGS

opencodemcpdevtoolsai-coding-agentopen-sourceproduct-update

DISCOVERED

1h ago

2026-07-28

PUBLISHED

2h ago

2026-07-28

RELEVANCE

7/ 10

AUTHOR

OpenCodeLog

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

NEWS54m ago

Gauntlet Loops Writes Horror Novel in Live Demo

An online post demonstrates expanding the use of Gauntlet Loops—an iterative AI agent workflow featuring dedicated critique loops—beyond game generation to draft a full horror novel. The live execution, hosted on Workbench, lets users observe the agentic writing process in real time as the system continuously refines and outputs long-form narrative content.

UPDATE1h ago

Model Context Protocol Receives Major Spec Update

The Model Context Protocol (MCP) project released a major update to its open specification and developer documentation platform. The update introduces refined protocol standards, updated transport and message pattern guidelines, and streamlined documentation for client features, enhancing how AI models connect with external data sources and tools.

MODEL1h ago

Fish Audio releases S2.1 Pro voice AI

Fish Audio has launched S2.1 Pro, a new open-weight voice AI model designed to challenge proprietary text-to-speech leaders like ElevenLabs. The model enables users to clone any voice using a short 15-second audio sample, combining high-fidelity audio synthesis with open weights for community access and fine-tuning.