MCP · conductor_ask (one-click in Claude Desktop / Cursor)

# Nexevo MCP server config — same JSON for Claude Desktop / Cursor; add an
# mcp-remote bridge entry. Config file locations:
#   Mac:    ~/Library/Application Support/Claude/claude_desktop_config.json
#   Win:    %APPDATA%\Claude\claude_desktop_config.json
#   Cursor: ~/.cursor/mcp.json  or  per-project .cursor/mcp.json

cat > claude_desktop_config.json <<'EOF'
{
  "mcpServers": {
    "nexevo": {
      "command": "npx",
      "args": [
        "-y", "mcp-remote",
        "https://api.nexevo.ai/v1/mcp",
        "--header", "Authorization: Bearer YOUR_NEXEVO_API_KEY"
      ]
    }
  }
}
EOF

# After saving, fully quit Claude Desktop (Cmd+Q, not close window) and reopen.
# The 🔌 icon at the input bottom-right should show nexevo · conductor_ask available.

# ── 3 typical scenarios (mode picked from natural language) ──
#
# Scenario 1 · Have another AI review your work (mode=chat):
#   You:    Push this JWT verify code to Conductor and let GPT-5 find bugs
#   Claude → conductor_ask({ mode: chat, prompt: "...", template: find_bug, judge_model: gpt-5 })
#
# Scenario 2 · Save a capsule (mode=save_memory):
#   You:    Save our decision about Clerk vs Auth0; tag: auth, decision
#   Claude → conductor_ask({ mode: save_memory, title: "Clerk vs Auth0", content: "...", tags: [...] })
#
# Scenario 3 · Recall a capsule (mode=search_memory):
#   You:    Find my prior capsule about webhook retries
#   Claude → conductor_ask({ mode: search_memory, query: "webhook retries", top_k: 3 })

# ── Direct curl (debug / your own agent integration) ──
curl https://api.nexevo.ai/v1/mcp \
  -H "Authorization: Bearer $NEXEVO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/call",
    "params": {
      "name": "conductor_ask",
      "arguments": {
        "mode":  "search_memory",
        "query": "webhook retry strategy",
        "top_k": 3
      }
    }
  }'

# Full doc: /docs/mcp