Mem0 MCP Serverofficial
A
licenseA
qualityF
maintenanceAn MCP server that integrates with mem0.ai to help users store, retrieve, and search coding preferences for more consistent programming practices.
2
9
658
Apache 2.0
Memstate AI - Agent Memory System
Officialby memstate-ai
Memstate AI is a hosted, versioned memory system for AI agents that lets you store, retrieve, search, and manage structured knowledge across sessions via the Model Context Protocol (MCP).
Store memories from markdown/text (
memstate_remember): Save task summaries, decisions, meeting notes, or any prose — the server automatically extracts keypaths, detects conflicts, and versions everythingSet specific key-value facts (
memstate_set): Store simple values at explicit hierarchical keypaths (e.g.config.database.port = 5432)Browse and retrieve memories (
memstate_get): List all projects, browse a project tree, retrieve a subtree by keypath, or fetch a single memory by ID — supports time-travel via revision numbersSemantic search (
memstate_search): Find memories by natural language meaning when you don't know the exact keypath, with optional filtering by project or categoryView version history (
memstate_history): Audit how a memory changed over time, see all past values with timestamps, and recover previous versionsSoft-delete memories (
memstate_delete/memstate_delete_project): Remove individual keypaths, subtrees, or entire projects while preserving full history via tombstones — deletions are reversibleConflict detection: Automatically identifies and preserves contradicting facts to maintain a reliable knowledge base
Constant token cost: Load summaries first and drill into details on demand — O(1) token usage regardless of memory scale
Cross-session continuity: Deterministic recall across sessions, compatible with Claude Desktop, Cursor, Cline, Windsurf, Kilo Code, and Roo Code
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., "@Memstate AI - Agent Memory Systemsearch my memories for the project deployment steps"
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.
Memstate AI - MCP
Versioned memory for AI agents. Store facts, detect conflicts, and track how decisions change over time — exposed as a hosted MCP server.
Why Memstate?
RAG (most other memory systems) | Memstate AI | |
Token usage per conversation | ~7,500 | ~1,500 |
Agent visibility | Black box | Full transparency |
Memory versioning | None | Full history |
Token growth as memories scale | O(n) | O(1) |
Infrastructure required | Yes | None — hosted SaaS |
Other memory systems dump everything into your context window and hope for the best. Memstate gives your agent a structured, versioned knowledge base it navigates precisely — load only what you need, know what changed, know when facts conflict.
Related MCP server: BuildAutomata Memory MCP Server
Benchmarks
We built an open-source benchmark suite that tests what actually matters for agent memory: can your system store facts, recall them accurately across sessions, detect conflicts when things change, and maintain context as a project evolves?
Head-to-Head: Memstate AI vs Mem0
Both systems were tested under identical conditions using the same agent (Claude Sonnet 4.6, temperature 0), the same scenarios, and the same scoring rubric.
Metric | Memstate AI | Mem0 | Winner |
Overall Score | 69.1 | 15.4 | Memstate |
Accuracy (fact recall) | 74.1 | 12.6 | Memstate |
Conflict Detection | 85.5 | 19.0 | Memstate |
Context Continuity | 63.7 | 10.1 | Memstate |
Token Efficiency | 22.3 | 30.6 | Mem0 |
Scoring weights: Accuracy 40%, Conflict Detection 25%, Context Continuity 25%, Token Efficiency 10%.
Per-Scenario Breakdown
The benchmark runs five real-world scenarios that simulate multi-session agent workflows:
Scenario | Memstate AI | Mem0 |
Web App Architecture Evolution | 43.2 | 55.6 |
Auth System Migration | 66.2 | 10.2 |
Database Schema Evolution | 72.7 | 7.0 |
API Versioning Conflicts | 86.5 | 0.9 |
Team Decision Reversal | 77.2 | 3.3 |
Mem0 won the first scenario (simple architecture tracking), but struggled severely on scenarios requiring contradiction handling, cross-session context, and decision reversal tracking — scoring near zero on three of five scenarios.
Why Memstate Wins
The benchmark reveals a fundamental architectural difference:
Mem0 uses embedding-based semantic search. Facts are chunked, embedded, and retrieved by similarity. This works for simple lookups but breaks down when:
Facts contradict earlier facts (the system can't distinguish current vs. outdated)
Precise recall is needed (embeddings return "similar" results, not exact ones)
Write-to-read latency matters (new memories take seconds to become searchable)
Memstate uses structured, versioned key-value storage. Every fact lives at an explicit keypath with a full version history. This means:
Conflict detection is built in — when a new fact contradicts an old one, the system knows and preserves both versions
Recall is deterministic — you get back exactly what was stored, not an approximate match
Cross-session continuity is reliable — the agent navigates a structured tree rather than hoping semantic search surfaces the right context
Token cost stays O(1) — the agent loads summaries first and drills into detail only when needed, instead of dumping all potentially-relevant embeddings into the context window
Fairness Notes
Both systems used the same agent model, temperature, and evaluation rubric
Mem0 was given a 10-second ingestion delay between writes and reads to account for its async embedding pipeline
Mem0 scores higher on token efficiency, but this metric should be read in context — lower token usage can simply reflect less information being returned. A system that retrieves incomplete or incorrect facts uses fewer tokens per response but may require more follow-up calls, ultimately costing more tokens to reach the same answer
The benchmark source code is included in this repository for full reproducibility
Mem0 may perform differently with custom configuration or a different embedding model
Quick Start
Get your API key at memstate.ai/dashboard, then add to your MCP client config:
{
"mcpServers": {
"memstate": {
"command": "npx",
"args": ["-y", "@memstate/mcp"],
"env": {
"MEMSTATE_API_KEY": "YOUR_API_KEY_HERE"
}
}
}
}No Docker. No database. No infrastructure. Running in 60 seconds.
Client Setup
Claude Desktop
Config location:
macOS:
~/Library/Application Support/Claude/claude_desktop_config.jsonWindows:
%APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"memstate": {
"command": "npx",
"args": ["-y", "@memstate/mcp"],
"env": { "MEMSTATE_API_KEY": "YOUR_API_KEY_HERE" }
}
}
}Claude Code
claude mcp add memstate npx @memstate/mcp -e MEMSTATE_API_KEY=YOUR_API_KEY_HERECursor
In Cursor Settings → MCP → Add Server — same JSON format as Claude Desktop above.
Cline / Windsurf / Kilo Code / Roo Code
All support the same stdio MCP config format. Add to your client's MCP settings file.
Core Tools
Tool | When to use |
| Store markdown, task summaries, decisions. Server extracts keypaths and detects conflicts automatically. Use for most writes. |
| Set a single keypath to a short value (e.g. |
| Browse all memories for a project or subtree. Use at the start of every task. |
| Semantic search by meaning when you don't know the exact keypath. |
| See how a piece of knowledge changed over time — full version chain. |
| Soft-delete a keypath. Creates a tombstone; full history is preserved. |
| Soft-delete an entire project and all its memories. |
How keypaths work
Memories are organized in hierarchical dot-notation:
project.my_app.database.schema
project.my_app.auth.provider
project.my_app.deploy.environmentKeypaths are auto-prefixed: keypath="database" with project_id="my_app" → project.my_app.database. Your agent can drill into exactly what it needs — no full-context dumps.
How It Works
Agent: memstate_remember(project_id="my_app", content="## Auth\nUsing SuperTokens...")
↓
Server extracts keypaths: [project.my_app.auth.provider, ...]
↓
Conflict detection: compare against existing memories at those keypaths
↓
New version stored — old version preserved in history chain
↓
Next session: memstate_get(project_id="my_app") → structured summaries only
↓
Agent drills into project.my_app.auth only when it needs auth detailsToken cost stays constant regardless of how many total memories exist.
Add to Your Agent Instructions
Copy into your AGENTS.md or system prompt:
## Memory (Memstate MCP)
### Before each task
- memstate_get(project_id="my_project") — browse existing knowledge
- memstate_search(query="topic", project_id="my_project") — find by meaning
### After each task
- memstate_remember(project_id="my_project", content="## Summary\n- ...", source="agent")
### Tool guide
- memstate_remember — markdown summaries, decisions, task results (preferred)
- memstate_set — single short values only (config flags, status)
- memstate_get — browse/retrieve before tasks
- memstate_search — semantic lookup when keypath unknown
- memstate_history — audit how knowledge evolved
- memstate_delete — remove outdated memories (history preserved)Environment Variables
Variable | Default | Description |
| (required) | API key from memstate.ai/dashboard |
|
| Override for self-hosted deployments |
Verify Your Connection
MEMSTATE_API_KEY=your_key npx @memstate/mcp --testPrints all available tools and confirms your API key works.
Built for AI agents that deserve to know what they know.
Available Tools
7 toolsmemstate_deleteA
Soft-delete a memory by keypath. Creates a tombstone version preserving full history. The memory can be un-deleted by setting a new value at the same keypath.
USE THIS WHEN: You need to remove outdated or incorrect memories, clean up a keypath subtree, or mark memories as no longer relevant. NOT FOR: Updating content (use memstate_set or memstate_remember to overwrite with new content instead).
memstate_delete(project_id="myapp", keypath="config.old_setting") memstate_delete(project_id="myapp", keypath="config", recursive=true) → deletes config and all children
History is preserved. Use memstate_history to see the deletion in the version chain. Keypath is auto-prefixed.
ParametersJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| project_id | Yes | Required. Project containing the memory. | |
| keypath | Yes | Required. Keypath to delete (auto-prefixed with 'project.{project_id}.'). | |
| recursive | No | If true, delete the entire keypath subtree. Default: false. |
Output Schema
ParametersJSON Schema
| Name | Required | Description |
|---|---|---|
| result | Yes |
TDQS
A4.5/5.0
Behavior4/5
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It effectively describes key behavioral traits: soft-deletion (not permanent), tombstone creation, history preservation, undeletion capability, and auto-prefixing behavior. It also references memstate_history for viewing deletions. The only minor gap is lack of explicit mention about permissions or error conditions.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Conciseness5/5
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is well-structured with clear sections, uses bold headings for emphasis, includes practical examples, and every sentence adds value. It's appropriately sized for a tool with important behavioral nuances and sibling distinctions.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Completeness5/5
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's complexity (soft-deletion with history preservation), the description provides comprehensive context. With no annotations but an output schema present, it covers behavioral aspects thoroughly while appropriately deferring return value details to the output schema. It addresses key sibling relationships and usage scenarios.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Parameters3/5
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so the baseline is 3. The description adds some value by providing concrete examples with project_id='myapp' and showing recursive behavior, but doesn't significantly enhance the parameter understanding beyond what the schema already documents clearly.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Purpose5/5
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the specific action ('soft-delete a memory by keypath') and distinguishes it from siblings by explaining it creates tombstone versions and preserves history. It explicitly contrasts with memstate_set and memstate_remember for updating content.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Usage Guidelines5/5
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description includes explicit 'USE THIS WHEN' and 'NOT FOR' sections that provide clear guidance on when to use this tool versus alternatives. It mentions specific scenarios (removing outdated memories, cleaning up subtrees) and names alternative tools for different use cases.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
memstate_delete_projectA
Soft-delete an entire project and all its memories. Creates tombstone versions for every memory, preserving full history. The project is hidden from listings but can be restored by creating a new project with the same ID.
USE THIS WHEN: You want to remove all memories for a project, e.g. cleaning up test data or decommissioning a project.
memstate_delete_project(project_id="old-project")
All memories get individual tombstone versions. History is preserved via memstate_history.
ParametersJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| project_id | Yes | Required. Project ID to soft-delete. |
Output Schema
ParametersJSON Schema
| Name | Required | Description |
|---|---|---|
| result | Yes |
TDQS
A4.6/5.0
Behavior4/5
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden and does so well. It discloses key behavioral traits: it's a soft-delete (not permanent), creates tombstone versions for all memories, preserves history, hides the project from listings, and can be restored by creating a new project with the same ID. It doesn't mention rate limits or auth needs, but covers the essential destructive nature and recovery mechanism adequately.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Conciseness4/5
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is appropriately sized and front-loaded with the core action. The 'USE THIS WHEN' section is helpful but slightly repetitive of the purpose. The example usage is concise and adds clarity. Minor improvements could tighten it, but overall it's efficient with minimal waste.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Completeness5/5
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's complexity (destructive operation with recovery), no annotations, and an output schema (which handles return values), the description is complete. It covers purpose, usage, behavior, and parameters adequately, providing all necessary context for an agent to invoke it correctly without over-explaining.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Parameters4/5
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema description coverage is 100%, so the baseline is 3. The description adds value by providing an example usage (memstate_delete_project(project_id='old-project')) which clarifies the parameter's role in context, though it doesn't add deep semantic details beyond what the schema already states ('Required. Project ID to soft-delete.').
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Purpose5/5
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the specific action ('soft-delete an entire project and all its memories') and distinguishes it from siblings like memstate_delete (which likely deletes individual memories) by specifying it affects the entire project. It explains the outcome (creates tombstone versions, preserves history, hides from listings) which clarifies the scope beyond just the name.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Usage Guidelines5/5
Does the description explain when to use this tool, when not to, or what alternatives exist?
It includes an explicit 'USE THIS WHEN' section that provides clear guidance: 'You want to remove all memories for a project, e.g. cleaning up test data or decommissioning a project.' This directly tells the agent when to use this tool versus alternatives, with practical examples.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
memstate_getA
Browse and retrieve memories. Use BEFORE starting tasks to fetch existing knowledge.
USE THIS WHEN: You need to read what's already stored — list projects, browse a project tree, or get content at a keypath. NOT FOR: Saving new content (use memstate_remember for markdown/summaries, memstate_set for one keypath value).
memstate_get() → list all projects memstate_get(project_id="myapp") → full project tree memstate_get(project_id="myapp", keypath="database") → subtree memstate_get(project_id="myapp", keypath="db", include_content=true) → with full content memstate_get(memory_id="mem_abc123") → single memory by ID
Keypath is auto-prefixed. Default recursive=true.
ParametersJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| project_id | No | Project to retrieve from. If omitted, lists all projects. | |
| keypath | No | Subkeypath within project. If omitted with project_id, returns the full project tree. | |
| memory_id | No | Alternative: get a single memory by its UUID | |
| recursive | No | Include subtree (default: true) | |
| include_content | No | If true, include full memory content in response (not just keypaths) | |
| at_revision | No | Optional revision number for time-travel queries | |
| content_limit | No | Max memories to hydrate when include_content=true (default: 200, max: 1000) |
Output Schema
ParametersJSON Schema
| Name | Required | Description |
|---|---|---|
| result | Yes |
TDQS
A4.5/5.0
Behavior4/5
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries full burden and adds valuable behavioral context. It explains that 'Keypath is auto-prefixed' and 'Default recursive=true', which are important behavioral traits not evident from the schema alone. It also provides example usage patterns showing different parameter combinations and their effects.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Conciseness5/5
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is well-structured with clear sections, bullet points for examples, and zero wasted text. Every sentence serves a purpose: establishing context, providing usage guidelines, showing examples, and noting important defaults. It's appropriately sized for a tool with 7 parameters and complex usage patterns.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Completeness5/5
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's complexity (7 parameters, multiple usage patterns) and the presence of an output schema, the description is complete. It covers purpose, guidelines, behavioral traits, and provides concrete examples. The output schema existence means the description doesn't need to explain return values, and it adequately addresses the tool's scope and limitations.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Parameters3/5
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so the baseline is 3. The description adds some value through the usage examples that show how parameters combine (e.g., project_id alone vs. with keypath, include_content=true), but doesn't provide significant additional semantic meaning beyond what's already documented in the schema descriptions.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Purpose5/5
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'Browse and retrieve memories' with specific verbs and resources. It distinguishes from siblings by explicitly contrasting with memstate_remember (for saving markdown/summaries) and memstate_set (for one keypath value), making the distinction clear.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Usage Guidelines5/5
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides explicit guidance: 'Use BEFORE starting tasks to fetch existing knowledge' and includes dedicated 'USE THIS WHEN:' and 'NOT FOR:' sections. It names specific alternatives (memstate_remember, memstate_set) and gives concrete scenarios for when to use this tool (list projects, browse project tree, get content at keypath).
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
memstate_historyA
View version history for a keypath or memory chain. Use when you need to see how a value changed over time.
USE THIS WHEN: Debugging past state, auditing changes, or recovering a previous value. NOT FOR: Reading current content (use memstate_get) or saving (use memstate_remember or memstate_set).
memstate_history(project_id="myapp", keypath="config.database.port") memstate_history(memory_id="mem_abc123")
Returns all versions with timestamps; latest is marked is_latest=true.
ParametersJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| project_id | No | Required with keypath | |
| keypath | No | Get history for this keypath | |
| memory_id | No | Or get history for a specific memory chain by ID |
Output Schema
ParametersJSON Schema
| Name | Required | Description |
|---|---|---|
| result | Yes |
TDQS
A4.7/5.0
Behavior4/5
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It effectively describes the tool's behavior: it returns 'all versions with timestamps' and marks the latest with 'is_latest=true.' However, it lacks details on potential limitations like rate limits, pagination, or error conditions, which would be helpful for a read operation with historical data.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Conciseness5/5
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is well-structured and front-loaded: it starts with the core purpose, followed by usage guidelines, examples, and return details. Every sentence adds value—no waste. The use of bold headers and code examples improves readability without unnecessary verbosity.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Completeness5/5
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's moderate complexity (3 parameters, no annotations, but with an output schema), the description is complete enough. It covers purpose, usage, parameters, and return values. Since an output schema exists, the description doesn't need to explain return values in detail, and it adequately addresses the context provided by sibling tools and schema coverage.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Parameters4/5
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has 100% description coverage, so the baseline is 3. The description adds value by clarifying parameter semantics: it explains that 'project_id' is 'Required with keypath' (though this is also in the schema) and provides example usage with both 'keypath' and 'memory_id', showing how they are used alternatively. This enhances understanding beyond the schema's basic descriptions.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Purpose5/5
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'View version history for a keypath or memory chain.' It specifies the verb ('view') and resource ('version history'), and distinguishes it from siblings by explicitly naming alternatives (memstate_get for current content, memstate_remember/memstate_set for saving). This is specific and avoids tautology.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Usage Guidelines5/5
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides explicit usage guidelines with dedicated sections: 'USE THIS WHEN:' lists debugging, auditing, and recovery scenarios, and 'NOT FOR:' explicitly names when not to use it (reading current content or saving) and specifies alternative tools (memstate_get, memstate_remember, memstate_set). This gives clear context and exclusions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
memstate_rememberA
Save markdown, task summaries, or any text. Server extracts keypaths and creates structured memories automatically. This is the PREFERRED way to save information.
USE THIS WHEN: Saving task summaries, meeting notes, docs, or any text with multiple facts. The server handles organization, conflict detection, and versioning. NOT FOR: Setting one specific keypath to a short value (e.g. config.port = "8080") — use memstate_set for that.
memstate_remember(project_id="myapp", content="## Task Summary\n- Added OAuth\n- Files: auth.go, middleware.go", source="agent") memstate_remember(project_id="myapp", content="Architecture decision: migrated to JWT tokens for session management")
Content limit: 100,000 chars. Processing is async (~15-18s); returns job_id immediately.
ParametersJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| project_id | Yes | Required. Project to store in (e.g. 'myapp'). Auto-creates if new. | |
| content | Yes | Markdown or text to remember (max 100,000 chars). Server extracts keypaths and creates structured memories automatically. | |
| source | No | Source type: agent, readme, docs, meeting, code | |
| context | No | Optional hint to guide keypath extraction |
Output Schema
ParametersJSON Schema
| Name | Required | Description |
|---|---|---|
| result | Yes |
TDQS
A4.5/5.0
Behavior4/5
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries full burden and does well by disclosing key behavioral traits: content limit (100,000 chars), async processing with timing (~15-18s), immediate job_id return, automatic project creation, and server-side organization/conflict detection/versioning. It doesn't mention error conditions or retry behavior, keeping it from a perfect score.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Conciseness5/5
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is efficiently structured with clear sections, front-loaded key information, and every sentence adds value. It uses bold headings for guidance, includes practical examples, and provides critical constraints without unnecessary elaboration.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Completeness5/5
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's complexity (async processing, automatic extraction), 100% schema coverage, and presence of an output schema, the description provides excellent context. It covers purpose, usage guidelines, behavioral traits, constraints, and examples, making it complete enough for an agent to understand when and how to use this tool effectively.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Parameters3/5
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so the baseline is 3. The description adds minimal parameter semantics beyond the schema: it reinforces content limits and provides example values in the code snippets, but doesn't significantly enhance understanding of parameters beyond what the schema already documents well.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Purpose5/5
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'Save markdown, task summaries, or any text' with server-side extraction of keypaths and structured memory creation. It explicitly distinguishes from sibling tools by stating this is the 'PREFERRED way to save information' and contrasting with memstate_set for specific keypath assignments.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Usage Guidelines5/5
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides explicit guidance with dedicated 'USE THIS WHEN:' and 'NOT FOR:' sections, listing specific use cases (task summaries, meeting notes, docs) and clear exclusions (setting one specific keypath to short values). It names the alternative tool (memstate_set) for excluded scenarios.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
memstate_searchA
Find memories by meaning (semantic search). Use BEFORE starting tasks when you don't know the exact keypath.
USE THIS WHEN: You want to find relevant memories by topic or meaning, not by exact keypath. NOT FOR: Saving content (use memstate_remember for markdown/summaries, memstate_set for one keypath value).
memstate_search(query="how is authentication configured", project_id="myapp") memstate_search(query="database connection settings", project_id="myapp")
Returns summaries with similarity scores. Use memstate_get(memory_id="...") to fetch full content of a result.
ParametersJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| query | Yes | Natural language search query. Leave empty to explore/list all memories ordered by keypath. | |
| project_id | No | Filter by project | |
| limit | No | Maximum results (default: 20, max: 100) | |
| categories | No | Filter by categories | |
| keypath_prefix | No | Filter by keypath prefix | |
| include_superseded | No | Include old versions |
Output Schema
ParametersJSON Schema
| Name | Required | Description |
|---|---|---|
| result | Yes |
TDQS
A4.4/5.0
Behavior4/5
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It effectively describes key behaviors: it's a search tool (implied read-only), returns summaries with similarity scores, and requires memstate_get to fetch full content. However, it doesn't mention potential limitations like rate limits or authentication needs, which could be relevant for a search tool.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Conciseness5/5
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is well-structured and front-loaded with the core purpose. Every sentence adds value: the initial statement, usage guidelines, examples, and instructions for fetching full content. It uses bold formatting effectively without being verbose, and there's no wasted text.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Completeness4/5
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's complexity (semantic search with 6 parameters), no annotations, but with a rich input schema (100% coverage) and output schema (confirmed present), the description is mostly complete. It explains the tool's purpose, usage, and result handling well. The main gap is lack of explicit behavioral details like error conditions or performance characteristics, but the output schema likely covers return values.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Parameters3/5
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so the schema already documents all 6 parameters thoroughly. The description adds minimal parameter semantics beyond the schema—it only mentions 'query' and 'project_id' in examples. This meets the baseline of 3 for high schema coverage, but doesn't provide additional meaningful context about parameters.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Purpose5/5
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'Find memories by meaning (semantic search).' It specifies the verb ('Find'), resource ('memories'), and method ('semantic search'), and distinguishes it from siblings by contrasting with exact keypath searches. The title being null doesn't detract from this clarity.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Usage Guidelines5/5
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides explicit guidance with 'USE THIS WHEN:' and 'NOT FOR:' sections, clearly stating when to use this tool (for semantic search when exact keypath is unknown) and when to use alternatives (memstate_remember for saving content, memstate_set for one keypath value). It also mentions using it 'BEFORE starting tasks' for context.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
memstate_setA
Set one keypath to a short value. Only for simple key=value facts like config, status, or version numbers.
USE THIS WHEN: You need to store ONE specific value at a known keypath (e.g. a port number, a status flag, a version string). NOT FOR: Task summaries, markdown, or text with multiple facts — use memstate_remember instead.
memstate_set(project_id="myapp", keypath="config.database.port", value="5432") memstate_set(project_id="myapp", keypath="status.deployment", value="production") memstate_set(project_id="myapp", keypath="version.current", value="2.1.0")
Value limit: 2,000 chars. If longer, use memstate_remember instead. Keypath is auto-prefixed.
ParametersJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| project_id | Yes | Required. Project to store in (e.g. 'myapp'). Auto-creates if new. | |
| keypath | Yes | Required. Hierarchical path like 'config.port' or 'status'. Auto-prefixed with 'project.{project_id}.' | |
| value | Yes | The value to store (max 2,000 chars). Short, simple values only — not markdown. | |
| category | No | Category: decision, preference, fact, task, context, requirement, note, code, learning | |
| topics | No | Optional additional topics/tags |
Output Schema
ParametersJSON Schema
| Name | Required | Description |
|---|---|---|
| result | Yes |
TDQS
A4.5/5.0
Behavior4/5
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries full burden and delivers well. It discloses key behavioral traits: value limit (2,000 chars), auto-prefixing of keypath, and that it's for storage/mutation (implied by 'set'). It doesn't mention error conditions, permissions, or rate limits, but provides substantial operational context beyond basic function.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Conciseness5/5
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is perfectly structured and concise. It opens with the core purpose, provides clear usage guidelines in dedicated sections, includes practical examples, and adds important constraints. Every sentence earns its place with no wasted words, and information is front-loaded appropriately.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Completeness5/5
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool's moderate complexity (storage operation with 5 parameters), 100% schema coverage, and presence of an output schema, the description is complete. It covers purpose, usage guidelines, constraints, and provides examples. The output schema means return values don't need explanation in the description.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Parameters3/5
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so the baseline is 3. The description adds minimal parameter semantics beyond the schema: it reinforces that value should be 'short, simple values only' and provides concrete examples of keypath usage. However, it doesn't explain the optional 'category' parameter's purpose or provide guidance on when to use topics.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Purpose5/5
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's purpose: 'Set one keypath to a short value' with specific examples of what to store (config, status, version numbers). It explicitly distinguishes from sibling memstate_remember by stating 'Only for simple key=value facts' and 'NOT FOR: Task summaries, markdown, or text with multiple facts.'
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Usage Guidelines5/5
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides explicit usage guidance with dedicated sections: 'USE THIS WHEN:' for appropriate scenarios (storing one specific value at known keypath) and 'NOT FOR:' for exclusions (task summaries, markdown, multi-fact text). It names the alternative tool (memstate_remember) and includes a value length threshold (2,000 chars) for when to switch tools.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
1 tool update
v1.0.2- Changed
memstate_search1 field changed- changed
Input schema / properties / query / descriptionPrevious value: -"Natural language search query"New value: +"Natural language search query. Leave empty to explore/list all memories ordered by keypath."
7 tool updates
v1.0.1- First observed
memstate_delete - First observed
memstate_delete_project - First observed
memstate_get - First observed
memstate_history - First observed
memstate_remember - First observed
memstate_search - First observed
memstate_set
TDQS
A4.7/5.0
Disambiguation5/5
Each tool has a clearly distinct purpose with minimal overlap. The descriptions explicitly differentiate tools (e.g., memstate_remember vs. memstate_set, memstate_get vs. memstate_search) and include 'USE THIS WHEN' and 'NOT FOR' sections that prevent confusion. The tools cover different aspects of memory management without redundancy.
Naming Consistency5/5
All tools follow a consistent 'memstate_verb' naming pattern (e.g., memstate_delete, memstate_get, memstate_search). This uniformity makes it easy to identify the server's tools and their functions at a glance, with no deviations in style or structure across the set.
Tool Count5/5
With 7 tools, the server is well-scoped for an agent memory system. Each tool serves a specific, necessary function (e.g., create, read, update, delete, search, history, project management), and there are no extraneous tools. The count aligns with the domain's complexity without being overwhelming or insufficient.
Completeness5/5
The tool set provides comprehensive coverage for memory management, including CRUD operations (memstate_remember/set for create/update, memstate_get for read, memstate_delete for delete), search (memstate_search), version history (memstate_history), and project-level management (memstate_delete_project). There are no obvious gaps; agents can handle full memory lifecycles and workflows effectively.
Maintenance
ActivityInactive
ResponsivenessNo issues
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Connectors
Persistent memory for AI agents. EU-hosted, privacy-first, hybrid recall, contradiction detection.
Hosted persistent memory with semantic search, importance and TTL for AI agents.
Persistent knowledge graph for AI-augmented teams. Store decisions, findings, and standing rules across agent sessions with semantic search and typed connections. Includes cross-session memory, audit trail, workspace isolation, and secret detection. Built for teams running agents that need to remember. Free until launch with team tier as default, anon trial available.
Universal memory for AI agents and tools. Save, organize and search context anywhere.
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceProvides AI agents with persistent, searchable memory that survives across conversations using semantic search, temporal versioning, and smart organization. Enables long-term context retention and cross-session continuity for AI assistants.14-
- FlicenseNot gradedqualityNot gradedmaintenanceEnables ChatGPT Desktop to save and manage conversation memories locally in SQLite, providing memory search, summarization, and markdown export capabilities to replace the built-in memory feature when in developer mode.-
- AlicenseNot gradedqualityCmaintenanceA local, persistent memory system for AI coding assistants that stores decisions, patterns, and session context via MCP tools. It enables cross-session memory management using SQLite and optional vector search without external dependencies or cloud storage.5864MIT
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/memstate-ai/memstate-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server