Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Deep Recall MCP Server Remember that I prefer Python over Java"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Your AI agent already thinks. We give it a memory.
Other memory systems intercept your conversations and run them through a separate LLM to decide what's worth remembering. That's like having a stranger take notes at your therapy session — they don't know what's significant to you.
Your agent IS an LLM. It already understands the conversation. Deep Recall gives it a memory layer with biological properties: memories that strengthen with use, fade when stale, catch their own contradictions, and self-organize into knowledge clusters. No extra LLM calls. No per-memory API costs. 41ms search.
pip install deeprecall-mcpRelated MCP server: genesys-memory
Sign up at deeprecall.dev/signup or use the API:
curl -X POST https://api.deeprecall.dev/v1/signup \
-H "Content-Type: application/json" \
-d '{"name": "Your Name", "email": "you@example.com", "password": "your-password"}'Save the api_key from the response — it's only shown once.
Add to ~/.claude/settings.json:
{
"mcpServers": {
"deeprecall": {
"command": "deeprecall-mcp",
"env": {
"DEEPRECALL_API_KEY": "ec_live_YOUR_KEY_HERE"
}
}
}
}Add to .cursor/mcp.json in your project root:
{
"mcpServers": {
"deeprecall": {
"command": "deeprecall-mcp",
"env": {
"DEEPRECALL_API_KEY": "ec_live_YOUR_KEY_HERE"
}
}
}
}Same JSON format in your MCP configuration file.
Your AI now has memory tools. Try saying:
"Remember that I prefer TypeScript over JavaScript"
"What do you know about me?"
"Search your memory for anything about our API architecture"
"Check if any of your memories contradict each other"
Two tools. That's it.
Tool | What it does |
| Find memories. Hybrid keyword + semantic, salience-weighted. |
| Store a memory. All biology runs automatically. |
Your agent searches early, remembers what matters. Behind the scenes, every store automatically:
Embeds for semantic search
Builds graph edges to related memories
Detects contradictions with existing knowledge
Resolves temporal changes ("moved to NYC" auto-supersedes "lives in SF")
Infers entity relationships from co-occurrence
Consolidates episode clusters into durable facts
Decays unused memories, strengthens recalled ones
No LLM calls. Pure biology in milliseconds. Two tools in your context window.
Deep Recall | Mem0 | Zep | Letta | |
Extra LLM calls | None | Required | Required | Required |
Search latency | 41ms | ~200ms | ~200ms | ~300ms |
Intelligent forgetting | ACT-R | No | No | No |
Hebbian reinforcement | Yes | No | No | No |
Contradiction detection | Yes | No | No | No |
Emotional context | Yes | No | No | No |
Agent decides what to store | Yes | No — LLM decides | No — LLM decides | Partial |
Plan | Price | Memories | Features |
Free | $0/mo | 10,000 | All core features, 30 req/min |
Builder | $19/mo | 100,000 | + topology, 120 req/min |
Pro | $49/mo | 1,000,000 | + emotional search, priority support |
Enterprise | $149/mo | 10,000,000 | + dedicated support, 3,000 req/min |
Website: deeprecall.dev
Quick Start: deeprecall.dev/quickstart
API Docs: api.deeprecall.dev/docs
Dashboard: api.deeprecall.dev/dashboard
npm SDK: @zappaidan/deeprecall
Email: aidan@deeprecall.dev
Built by Aidan Poole & Thomas.
Store a memory. Behind the scenes: embeds for semantic search, builds graph edges to related memories, detects contradictions, auto-resolves temporal changes, infers entity relationships, and periodically consolidates episode clusters into durable facts. All biology runs automatically — just store what matters.
| Name | Required | Description | Default |
|---|---|---|---|
| content | Yes | The memory to store | |
| person | No | Who this memory is about | |
| kind | No | Memory type | fact |
| salience | No | Importance 0-1. Higher resists decay longer. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the full burden. It details internal processes like embedding, graph edges, contradiction detection, temporal resolution, and consolidation, which provides substantial transparency about side effects and automatic behaviors.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is relatively concise but includes a long list of internal processes that may not be essential for an agent to know. It could be more streamlined by focusing on the core function.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
The description does not mention return values, side effects like success/failure responses, or prerequisites (e.g., authentication). Given no output schema, this omission is significant for an agent to understand the tool's full behavior.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100%, so baseline is 3. The description adds no additional meaning to the parameters beyond what the schema already provides, such as the enum for 'kind' or the range for 'salience'.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states 'Store a memory.' as the core action, with a specific verb and resource. It also distinguishes from the sibling tool 'deeprecall_search' which is for searching, not storing.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies usage by stating 'just store what matters' and contrasts with the sibling 'deeprecall_search', providing implicit context. However, it does not explicitly state when not to use or provide alternative guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Search memories. Hybrid keyword + semantic search, ranked by salience. Faded memories rank lower. 'outdoor activities' finds 'loves hiking'.
| Name | Required | Description | Default |
|---|---|---|---|
| query | Yes | Natural language search query | |
| person | No | Filter by person | |
| limit | No | Max results |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, so the description carries the full burden. It discloses that faded memories rank lower, which is a behavioral trait, but does not mention auth needs, rate limits, or return format.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise with three sentences, front-loaded with key information, and includes an illustrative example. Every sentence adds value without waste.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no output schema, the description could be more complete about what is returned. It covers key behavioral aspects like fading memories but lacks details on return format or result handling.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100% (all parameters have descriptions). The description adds context about the search functionality but does not add specific meaning to individual parameters beyond what the schema provides.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description uses specific verbs ('Search memories') and describes the hybrid keyword+semantic search and ranking by salience, distinguishing it from the sibling tool deeprecall_remember which likely stores memories.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No explicit guidance on when to use vs. alternatives, but the description implies it is for searching while the sibling is for remembering. No exclusions or when-not-to-use are stated.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
v0.4.0deeprecall_rememberdeeprecall_searchThe two tools are clearly distinct: one stores a memory, the other searches memories. There is no overlap or ambiguity in their purposes.
Both tools follow a consistent verb_noun pattern with the 'deeprecall_' prefix, using 'remember' and 'search' as clear action verbs.
With only 2 tools, the server feels minimal. While it covers basic store and search, it lacks additional tools for management, making the count borderline for a memory system.
The server lacks common operations like delete, update, or list memories. Although it claims automatic resolution, agents may need explicit lifecycle tools, leaving significant gaps.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Persistent memory and knowledge management for AI agents with semantic search and 50+ tools.
Persistent memory for AI agents. EU-hosted, privacy-first, hybrid recall, contradiction detection.
Universal memory for AI agents and tools. Save, organize and search context anywhere.
Universal memory runtime for AI agents — episodic, semantic, and procedural memory.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Relic-Studios/deeprecall-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
