bulk_delete
Delete multiple memories at once by their IDs. Maximum 100 IDs per call. This is permanent.
LABSCO SUMMARY
The store does not just hold memories — it scores them, ages them, and tells you which ones nothing has ever recalled.
Most of the surface is upkeep rather than storage: get_analytics separates the memories that keep getting recalled from those that never have, consolidate_memories folds clusters into synthesised facts while archiving the originals with version history, and feedback_memory feeds the ranking back. The heavier tools are model passes rather than lookups — extract_memories, ingest_document, consolidate_memories, generate_skills and extract_principles all run over your content, and get_usage reports extractions this month against a tier, so that work is metered. Rules are the deliberate exception to the ranking: they are injected into every conversation regardless of score or decay.
What it is
A hosted memory service behind 37 tools, holding facts, rules, tasks and corrections for a project and returning them to later sessions.
What you get
- Recall at session start and at the moment before a suggestion: get_project_context loads the top memories at session start, preflight runs a keyword search before a suggestion is made, and recall_memories runs semantic search filtered by type, category, tag, scope or minimum importance.
- Memories that outrank scoring: save_rule stores a directive injected into every conversation, bypassing scoring and decay, with list_rules and delete_rule to keep the set honest.
- Capture without dictation: extract_memories reads a message-and-response exchange and identifies what is worth keeping, ingest_document chunks a README or API spec into memories, and save_session_summary records what a session settled.
- A safety net over deletion: delete_memory moves an item to the trash where restore_memory and list_trash can recover it for 30 days, and bulk_delete and bulk_update take up to 100 and 50 items per call.
- Structure across memories: link_memories connects two entries with a named relationship, get_memory_links walks the graph, get_memory_versions shows how a fact changed, and list_tags reports the threads in use.
- Upkeep tools that name their own targets: get_analytics reports most-recalled, never-recalled and stale entries, get_stale_memories lists candidates by age, consolidate_memories synthesises clusters of three or more memories sharing a subject, and promote_memory lifts a project fact to global scope.
- Continuity beyond facts: save_task, get_tasks and update_task carry pending and in_progress work across sessions, set_reminder surfaces a note at the next session start, and get_file_changes diffs the repository against the snapshot from last time.
Requirements
A CogmemAi account and its key in COGMEMAI_API_KEY. get_file_changes compares against git state, so it wants the project under version control.
Setup effort
One command plus a key — npx cogmemai-mcp setup, then supply credentials
USE CASES
01
Picking a project back up after two weeks on something else
Load the project's top memories, read the tasks that were left open, and see which files moved while you were away.
02
Stopping the assistant from re-proposing something the team already rejected
Check prior context before the suggestion is made, and record the corrections so the same wrong turn surfaces automatically next time.
03
Getting a new codebase's documentation into project memory on day one
Feed the README and architecture docs in as memories, let an exchange be mined for the facts it settled, and check what came out.
04
Cleaning up a memory store that has grown noisy after months of sessions
Find what is never recalled or has gone stale, fold overlapping entries into single facts, and clear the rest in one call.
05
Making a standing instruction that every future session has to follow
Save it as a rule rather than a memory, so it is injected into every conversation instead of competing for recall, and review the set later.
37 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
bulk_update
Update multiple memories at once. Each item needs a memory_id and fields to update. Maximum 50 items per call.
consolidate_memories
Consolidate related memories into fewer, richer memories. Finds clusters of memories sharing the same subject (3+ memories required), then uses AI to synthesize each cluster into 1-2 comprehensive facts. Originals are archived (not deleted) with full version history. Use dry_run=true to preview without making changes. Great for cleaning up memory clutter after many sessions.
delete_memory
Move a memory to the trash by its ID. Trashed memories are hidden from recall and restorable with restore_memory for 30 days, then purged permanently.
delete_rule
Delete a mandatory rule by its memory ID. Use list_rules to find the ID first.
export_memories
Export all memories as JSON. Use this to back up memories or transfer them to another project.
extract_memories
Extract memories from a conversation exchange using AI. Send the developer message and assistant response, and the server identifies facts worth remembering (architecture decisions, preferences, bug fixes, etc.).
extract_principles
Analyze memory clusters and extract underlying factual principles or patterns. Principles are observations about your project — "this codebase tends to have X" — not behavioral instructions (those are skills). Use dry_run to preview candidates first.
feedback_memory
Signal whether a recalled memory was useful or irrelevant. Helps improve future recall quality over time. Use after recalling memories to indicate which were helpful vs noise.
generate_skills
Manually trigger skill generation from your corrections, preferences, and patterns. Skills are behavioral directives that auto-improve how the AI works with you. CogmemAi also generates skills automatically when enough evidence accumulates — this tool lets you trigger it manually or preview candidates.
get_analytics
Get a memory health dashboard with insights: most recalled memories, never-recalled memories, stale memories, growth trends, and breakdowns by type and category. Use this to identify cleanup opportunities and understand memory usage patterns.
get_file_changes
Show what files changed since your last session. Compares the current git state to a snapshot saved when your previous session ended. Helps you understand what happened between sessions.
get_memory_links
View all memories linked to a specific memory. Returns the relationship type and full memory details for each connection. Use this to explore the knowledge graph around a memory.
get_memory_versions
View the edit history of a memory. Shows all previous versions with timestamps and what changed. Useful for understanding how a decision or fact evolved over time.
get_project_context
Load top memories for the current project plus relevant global memories. Use at the start of a session to get full context from previous sessions. Optionally pass context to get memories most relevant to your current task.
get_stale_memories
Find memories that may be outdated based on age and access patterns. Returns memories that haven't been recalled or updated recently, so you can review, update, or delete them.
get_tasks
Get tasks for the current project. Returns tasks filtered by status — defaults to showing pending and in_progress tasks. Use at session start to pick up where you left off.
get_usage
Get current usage statistics — memory count, extractions this month, tier info, projects.
import_memories
Bulk import memories from a JSON array. Each memory needs at minimum a content field. Deduplication is applied automatically.
ingest_document
Extract memories from a document by splitting it into chunks and processing each one. Great for onboarding — feed in READMEs, architecture docs, or API specs to quickly build project context.
link_memories
Connect two related memories with a named relationship. Use this to build a knowledge graph — e.g., linking a bug fix to the architecture decision that caused it, or connecting a preference to the pattern it led to.
list_memories
List stored memories with optional filters by type, category, scope, or project.
list_rules
List all mandatory rules for the current project and/or globally. Rules are always-on memories that surface in every session.
list_tags
List all tags in use across your memories, with counts. Use this to see what threads/groups exist and find related memories by tag.
list_trash
List trashed (deleted) memories, newest first, with days remaining before each is purged permanently. Restore any of them with restore_memory.
preflight
Proactive recall. Call this BEFORE making any suggestion, recommendation, or action plan. Sends the topic to a fast keyword search and returns any relevant prior context: previous attempts, decisions, contacts, evaluations. This prevents suggesting things that were already tried, rejected, or completed. Fast and cheap, use liberally.
promote_memory
Promote a project-scoped memory to global scope so it applies across all projects. Use this when you discover a preference or pattern that should be universal — e.g., "user prefers tabs over spaces" or "always use Bun instead of npm".
recall_memories
Search stored memories using semantic search. Returns memories ranked by relevance, importance, and recency. Use this to find relevant context from past sessions.
restore_memory
Restore a trashed (deleted) memory by its ID. Works within 30 days of deletion, before the trash is purged. Use list_trash to find restorable memories.
save_correction
Save a correction pattern — what went wrong and what the right approach is. These are surfaced automatically when similar situations arise in future sessions, helping avoid repeated mistakes.
save_memory
Store a developer memory (fact, preference, decision, architecture detail). Memories persist across all Claude Code sessions and are available in future conversations.
save_rule
Save a mandatory rule that will ALWAYS be followed in every session. Rules bypass scoring and decay — they are injected into every conversation, every time. Use for absolute requirements like "NEVER do X" or "ALWAYS do Y".
save_session_summary
Save a summary of the current coding session. Captures what was accomplished, decisions made, and next steps. Stored as a session_summary memory for future reference.
save_task
Create a task that persists across sessions. Tasks are tracked with status (pending, in_progress, done, blocked) and priority (high, medium, low). Use this to maintain continuity on multi-session work.
set_reminder
Set a reminder that surfaces automatically at the start of your next session. Use for follow-ups, things to check, or deferred work. Reminders auto-archive after being shown.
update_memory
Update an existing memory's content, importance, or scope.
update_task
Update a task's status, title, description, or priority. Use this to mark tasks as in_progress, done, or blocked as you work.
BEFORE YOU INSTALL
What to know first
Tools you get
37 tools when we connected. restore_memory and list_trash, the undo path for delete_memory and bulk_delete, are available but not in the documentation's list of 35.
Requires
- npx on PATH
- COGMEMAI_API_KEY in your environment
CONTEXT COST
Moderate Standing cost
Rated by the size of the 37 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
npx
TERMINAL
npx cogmemai-mcp setupHosted endpoint
ENDPOINT URL
https://hifriendbot.com/mcp/Add this hosted endpoint to your MCP client using the credentials listed under Requirements.
MCP Clients
Claude Code
CLAUDE CODE · JSON
{
"mcpServers": {
"cogmemai": {
"command": "cogmemai-mcp",
"env": {
"COGMEMAI_API_KEY": "<YOUR_COGMEMAI_API_KEY>"
}
}
}
}The bare cogmemai-mcp command needs the package installed globally: npx cogmemai-mcp setup does that for you, or run npm install -g cogmemai-mcp.
Cursor, Windsurf, Cline
JSON
{
"mcpServers": {
"cogmemai": {
"command": "npx",
"args": [
"-y",
"cogmemai-mcp"
],
"env": {
"COGMEMAI_API_KEY": "<YOUR_COGMEMAI_API_KEY>"
}
}
}
}Continue
CONTINUE · JSON
{
"mcpServers": [
{
"name": "cogmemai",
"command": "npx",
"args": [
"-y",
"cogmemai-mcp"
],
"env": {
"COGMEMAI_API_KEY": "<YOUR_COGMEMAI_API_KEY>"
}
}
]
}Any MCP client — remote
Endpoint: https://hifriendbot.com/mcp/
Auth: Bearer token — your cm_ API key
