Neural Memory — Configuration

View and modify neural memory settings.

Available actions

• view: Show current configuration as JSON
• set_mode: Change indexing mode
• ast_only — Fast, local-only, no API cost. Uses AST heuristics for summaries.
• api_only — Rich AI summaries via Claude. Sends code to API.
• both — Heuristic base + API enrichment for important nodes (default).
• add_exclude: Add a glob pattern to exclude files (e.g., **/tests/**)
• add_redaction_pattern: Add a custom regex for secret detection
• set_staleness_threshold: How many commits behind before the agent warns you (default: 5)

How to call

Via MCP tool (neural-memory configured as MCP server in Claude Code):

Tool: neural_config
{ "action": "view" }

{ "action": "set_mode", "value": "ast_only" }

Via Python (working directly in the project):

import asyncio
from neural_memory.server import neural_config, ConfigInput

# View config
asyncio.run(neural_config(ConfigInput(action="view")))

# Set indexing mode
asyncio.run(neural_config(ConfigInput(action="set_mode", value="ast_only")))

# Add an exclusion pattern
asyncio.run(neural_config(ConfigInput(action="add_exclude", value="**/tests/**")))

# Set staleness threshold
asyncio.run(neural_config(ConfigInput(action="set_staleness_threshold", value="10")))

Parameters

Data residency

If your code must stay local, use set_mode with ast_only. No code will leave your machine.