Complete API Reference
Tutorial guide to all 50 MCP tools with real-world examples and workflows
Introduction
JauMemory provides 50 MCP tools that combine persistent memory, multi-agent coordination, an encrypted credential vault, and a tool/skill registry for chaining external APIs into reusable workflows. This guide walks through each tool and shows how they connect into end-to-end automations.
What You'll Learn
- How to use each of the 50 tools effectively
- Real-world examples and workflows
- How tools connect and build on each other
- Best practices for different use cases
Which Tool When?
Fifty tools sounds like a lot, but you'll only ever reach for the category that matches the moment. Core memory is your daily bread: remember what matters, recall it later, update and forget as life moves on. Collections earn their keep when memories outgrow one pile and you want a folder per project, trip, or client. Agents are for when multiple AI sessions work as a team and each needs its own identity, memory, and error-learning. The vault lets you store an API key once, encrypted, so tools and skills can use it without ever showing it to anyone again. The tool registry wraps external APIs as callable tools, skills chain those tools into reusable workflows, and scheduling runs them on a cron so the memory acts without you. Analysis is your rear-view mirror: patterns, statistics, and consolidation of duplicates into insights. Authentication and discovery are the plumbing: log in once with a browser approval, and let get_guide teach any new agent the ropes. When in doubt, start with remember and recall; everything else can wait until the day it obviously can't.
How It All Connects
JauMemory's 50 tools work together as a connected ecosystem of memory, agents, credentials, tools, and skills:
The Memory + Skill Lifecycle
- Authenticate → Use
mcp_login+mcp_authenticateonce, credentials cache locally - Store → Use
rememberto create memories; encryption is per-user AES-256-GCM - Organize → Add to
collectionsfor grouping; analyze withanalyzeandmemory_stats - Retrieve → Use
recallfor search (keyword/semantic/hybrid) - Vault → Store API keys with
vault_store; values are write-only and never returned - Tools → Register HTTP endpoints with
tool_create, thentool_callwith auto credential injection - Skills → Chain tools into pipelines with
skill_create, run viaskill_execute - Schedule → Run skills on cron with
skill_schedule; monitor withskill_tasks_pending - Coordinate → Multi-agent workflows via
agent_collaborationand per-agent memory - Consolidate → Merge similar memories with
consolidateto reduce redundancy
Tool Categories Explained
Authentication
3-step browser-approval login flow. Credentials cache locally so subsequent runs resume without re-auth.
Core Memory
The foundation - store, retrieve, update, delete, analyze, and consolidate memories. Start here.
Discovery
Tool-catalog search and fetch (used by ChatGPT/OAuth MCP clients), plus get_guide onboarding topics and personas.
Agents
Create specialized AI agents with personalities, memories, error-learning, reflections, and collaboration.
Collections
Organize related memories into groups for easy management and consolidation into summaries.
Vault
Encrypted credential storage. AES-256-GCM at rest, write-only values, masked-only readback.
Tool Registry
Register HTTP/REST/GraphQL/webhook endpoints as callable tools. Vault credentials auto-inject.
Skill Registry
Chain tools into pipelines with conditional steps, LLM transforms, and trigger phrases.
Toolkit & Execution
Unified search across tools+skills, plus end-to-end skill execution with auto credential injection.
Berrry Integration
One-call register or create-and-register Berrry apps as JauMemory tools via NOMCP.
Scheduling & Tasks
Run skills on cron schedules, monitor execution history, retry failures, resume LLM-paused steps.
Quick Reference Table
| Tool | Purpose | Common Use Cases |
|---|---|---|
| Authentication (3) | ||
mcp_login | Initiate browser-approval auth flow | First-time setup |
mcp_authenticate | Complete login with the auth code from browser | After clicking Approve |
mcp_logout | Clear local creds and revoke server session | Switching accounts, security |
| Core Memory (7) | ||
remember | Store new memories (encrypted per-user) | Logging tasks, errors, insights |
recall | Search memories (keyword/semantic/hybrid) | Finding past solutions, context |
forget | Delete a specific memory | Removing outdated info |
update | Modify content or importance of a memory | Marking tasks done, corrections |
analyze | Pattern analysis over a time window | Daily/weekly reviews |
consolidate | Merge similar memories into insights | Reducing redundancy |
memory_stats | Filtered statistics: types, tags, importance | Tag frequency, distributions |
| Discovery (3) | ||
search | Search JauMemory tool catalog by query | Discoverability for OAuth/ChatGPT clients |
fetch | Fetch full docs for a specific tool by ID | Inspecting tool params at runtime |
get_guide | Onboarding guides: topic index, topics, personas, search | Agent self-onboarding, system prompts |
| Agents (7) | ||
create_agent | Create specialized AI agents | Frontend dev, backend dev, reviewers |
list_agents | List agents with optional status filter | Inventory, debugging |
agent_memory | Link memories to agents or recall by agent | Per-agent context |
agent_error_learning | 2-strike protocol for error learning | Teaching agents from mistakes |
agent_reflection | Capture learning/mistake/success reflections | Continuous improvement |
update_agent_name | Rename agent (Long Name:shortname) | Role transitions |
agent_collaboration | Coordinate multi-agent work | Pair programming, code review |
| Collections (8) | ||
create_collection | Create named groups of memories | Project folders, topic groups |
list_collections | List all your collections | Browsing |
get_collection | Get collection + member memories | Reviewing a sprint |
add_to_collection | Add memory to collection | Organizing related memories |
remove_from_collection | Remove memory (memory itself preserved) | Re-categorizing |
update_collection | Rename or re-describe a collection | Curating |
delete_collection | Delete collection (memories preserved) | Cleanup |
consolidate_collection | Summarize collection into one memory | Sprint summaries, power memories |
| Vault (3) | ||
vault_store | Store API credential (encrypted, write-only) | OpenAI, Stripe, GitHub keys |
vault_list | List credentials with masked values | Auditing what's stored |
vault_rotate | Replace secret value of a credential | Key rotation, compromise response |
| Tool Registry (5) | ||
tool_create | Register HTTP/REST/GraphQL/webhook tool | Wrapping external APIs |
tool_update | Update tool config (partial) | Endpoint changes, deactivation |
tool_call | Execute tool by slug, auto-inject creds | One-shot API calls |
tool_list | List/filter registered tools | Discovery, search |
tool_render | Render tool as markdown (creds redacted) | Documentation, sharing |
| Skill Registry (3) | ||
skill_create | Create skill workflow chaining tools | Multi-step automations |
skill_list | List/filter skills | Discovery |
skill_render | Render skill + linked tools as markdown | Documentation |
| Toolkit & Execution (2) | ||
toolkit_search | Unified search across tools and skills | Cross-cutting discovery |
skill_execute | Run skill end-to-end with credential injection | Production automations |
| Berrry Integration (2) | ||
berrry_register_tool | Register an existing Berrry app as a tool | One-call wrap of a Berrry app |
berrry_create_tool | Create new Berrry app and register in one step | Bootstrapping new apps |
| Scheduling & Tasks (7) | ||
skill_schedule | Schedule skill via cron (max 20/user, min 60s) | Daily reports, periodic syncs |
skill_schedule_list | List scheduled runs with status filters | Monitoring |
skill_schedule_cancel | Cancel (soft-delete) a scheduled run | Stopping a job |
skill_schedule_retrigger | Re-trigger failed/completed schedule | Recovery |
skill_tasks_pending | List paused + recently failed executions | Action queue |
skill_task_retrigger | Retrigger a failed task (alias) | Recovery |
skill_tasks_list | List all execution logs | History, debugging |
Core Memory Operations
These 7 tools form the foundation of JauMemory. Master these first, then explore agents, vault, tools, skills, and scheduling.