DoboCache for VS Code
io.github.dobo123123/dobocache
Local proxy + MCP server: keeps LLM prompt-cache prefixes stable so caching hits.
client:VS Code
transport:stdio
runtime:npm
Install DoboCache in VS Code
.vscode/mcp.json
{
"inputs": [
{
"type": "promptString",
"id": "dobocache-log-dir",
"description": "DOBOCACHE_LOG_DIR",
"password": true
},
{
"type": "promptString",
"id": "dobocache-whitebox",
"description": "DOBOCACHE_WHITEBOX",
"password": true
},
{
"type": "promptString",
"id": "dobocache-hit-price-per-m",
"description": "DOBOCACHE_HIT_PRICE_PER_M",
"password": true
},
{
"type": "promptString",
"id": "dobocache-miss-price-per-m",
"description": "DOBOCACHE_MISS_PRICE_PER_M",
"password": true
}
],
"servers": {
"dobocache": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"dobocache"
],
"env": {
"DOBOCACHE_LOG_DIR": "${input:dobocache-log-dir}",
"DOBOCACHE_WHITEBOX": "${input:dobocache-whitebox}",
"DOBOCACHE_HIT_PRICE_PER_M": "${input:dobocache-hit-price-per-m}",
"DOBOCACHE_MISS_PRICE_PER_M": "${input:dobocache-miss-price-per-m}"
}
}
}
}VS Code prompts for each input the first time the server starts, so no secret is written to the file. VS Code docs