diff --git a/README.md b/README.md
index 87c630e..dd0ee83 100644
--- a/README.md
+++ b/README.md
@@ -96,6 +96,7 @@ OpenVisio speaks the **Model Context Protocol** (the open standard from
 | **Codex** | OpenAI |
 | **Windsurf** | Codeium |
 | **GitHub Copilot** (Agent Mode / VS Code MCP) | GitHub / Microsoft |
+| **opencode** | SST |
 | **Cline** | open source |
 | **Continue** | Continue.dev |
 | **Zed** | Zed Industries |
@@ -106,6 +107,23 @@ OpenVisio speaks the **Model Context Protocol** (the open standard from
 > MIT-licensed tool and is not affiliated with or endorsed by these companies — it
 > simply implements the open Model Context Protocol they support.
 
+For clients that aren't auto-configured, add the `openvisio` server by hand. For
+example, [**opencode**](https://opencode.ai) reads `~/.config/opencode/opencode.json`
+(or a project-local `opencode.json`):
+
+```json
+{
+  "$schema": "https://opencode.ai/config.json",
+  "mcp": {
+    "openvisio": {
+      "type": "local",
+      "command": ["openvisio", "mcp", ".", "--watch"],
+      "enabled": true
+    }
+  }
+}
+```
+
 ### Why token cost matters whatever model you run
 
 Whether your agent calls **Anthropic Claude** (Opus, Sonnet, Haiku), **OpenAI**
@@ -312,7 +330,7 @@ Contributions welcome — see [CONTRIBUTING.md](CONTRIBUTING.md).
 
 MCP server · Model Context Protocol · Anthropic Claude · Claude Code · Cursor
 (Anysphere) · OpenAI Codex · GPT · Google Gemini · GitHub Copilot · Microsoft
-VS Code · Windsurf · Codeium · Cline · Continue · Zed · JetBrains · Meta Llama ·
+VS Code · Windsurf · Codeium · Cline · Continue · Zed · JetBrains · opencode (SST) · Meta Llama ·
 Mistral · DeepSeek · reduce token usage · save LLM tokens · lower API cost ·
 token-efficient context · context engineering · code graph · code knowledge graph ·
 codebase visualization · 3D code map · dependency graph · call graph · import graph ·
diff --git a/mcp/README.md b/mcp/README.md
index 0c7ab63..42b182b 100644
--- a/mcp/README.md
+++ b/mcp/README.md
@@ -178,6 +178,20 @@ command = "openvisio"
 args = ["mcp", ".", "--watch"]
 ```
 
+**opencode** — `~/.config/opencode/opencode.json` (or project-local `opencode.json`):
+```json
+{
+  "$schema": "https://opencode.ai/config.json",
+  "mcp": {
+    "openvisio": {
+      "type": "local",
+      "command": ["openvisio", "mcp", ".", "--watch"],
+      "enabled": true
+    }
+  }
+}
+```
+
 ### Troubleshooting: "MCP server failed to connect" (nvm + GUI editors)
 
 If you installed Node with **nvm** and your MCP client is a **GUI app** (Cursor,
diff --git a/mcp/package.json b/mcp/package.json
index c9b55f3..9b2cd4f 100644
--- a/mcp/package.json
+++ b/mcp/package.json
@@ -1,6 +1,6 @@
 {
   "name": "openvisio",
-  "version": "0.2.4",
+  "version": "0.2.5",
   "description": "Cut your AI coding agent's token usage. MCP server + CLI that turns any codebase into a deterministic code graph and serves Claude Code, Cursor, Codex & Windsurf a ranked, token-budgeted query surface — agents query structure (symbols, imports, call edges, path:line anchors) instead of crawling files. ~30x fewer exploration tokens. Local-first, read-only, no network, no embeddings.",
   "type": "module",
   "license": "MIT",