このプロジェクトは、Omi APIと連携するためのモデルコンテキストプロトコル（MCP）サーバーを提供します。このサーバーは、会話や思い出の読み取り、および新しい会話や思い出の作成のためのツールを提供します。
リポジトリをクローンする
npm installで依存関係をインストールする
次の変数を含む.envファイルを作成します。
API_KEY=your_api_key
APP_ID=your_app_idRelated MCP server: Omi Memories MCP Server
Smithery経由で Claude Desktop 用の Omi MCP Server を自動的にインストールするには:
npx -y @smithery/cli install @fourcolors/omi-mcp --client claudenpm run buildnpm run startホットリロードを使用した開発の場合:
npm run devMCPサーバーとやり取りするためのシンプルなテストクライアントが含まれています。プロジェクトをビルドしたら、以下を実行してください。
npm run testまたは直接:
./test-mcp-client.jsMCPサーバーが起動し、利用可能なツールをテストするための対話型メニューが表示されます。テストクライアントは、すべての操作にデフォルトのテストユーザーID（ test-user-123 ）を使用します。
ビルド ディレクトリをクリーンアップして最初から再構築するには:
npm run rebuildこの MCP サーバーを Anthropic Console または API 経由で Claude で使用するには:
MCP サーバーをローカルで起動します。
npm run startClaude 会話を設定するときは、MCP 接続を構成します。
{
"mcp_config": {
"transports": [
{
"type": "stdio",
"executable": {
"path": "/path/to/your/omi-mcp-local/dist/index.js",
"args": []
}
}
]
}
}クロードへのプロンプトの例:
Please fetch the latest 5 conversations for user "user123" using the Omi API.Claude は MCP を使用してread_omi_conversationsツールを実行します。
{
"id": "req-1",
"type": "request",
"method": "tools.read_omi_conversations",
"params": {
"user_id": "user123",
"limit": 5
}
}この MCP サーバーを Cursor で使用するには:
ターミナルで MCP サーバーを起動します。
npm run startカーソルで、「設定」>「拡張機能」>「MCPサーバー」に移動します。
次の設定で新しい MCP サーバーを追加します。
名前: Omi API
URL: stdio:/path/to/your/omi-mcp-local/dist/index.js
サーバーを有効にする
これで、カーソル内でOmiツールを直接使用できるようになりました。例:
@Omi API Please fetch memories for user "user123" and summarize them.カーソルは MCP サーバーと通信して、必要な API 呼び出しを実行します。
MCP サーバーは次のツールを提供します。
オプションのフィルターを使用して、特定のユーザーの Omi から会話を取得します。
パラメータ:
user_id (文字列): 会話を取得するユーザーID
limit （数値、オプション）：返される会話の最大数
offset （数値、オプション）：ページ区切りでスキップする会話の数
include_discarded (ブール値、オプション): 破棄された会話を含めるかどうか
statuses （文字列、オプション）：会話をフィルタリングするためのステータスのコンマ区切りリスト
特定のユーザーの Omi から思い出を取得します。
パラメータ:
user_id (文字列): 思い出を取得するユーザーID
limit (数値、オプション): 返されるメモリの最大数
offset （数値、オプション）：ページ区切りでスキップするメモリの数
特定のユーザー向けに Omi で新しい会話を作成します。
パラメータ:
text （文字列）：会話の全文
user_id (文字列): 会話を作成するユーザーID
text_source (文字列): テキストコンテンツのソース (オプション: "audio_transcript", "message", "other_text")
started_at (文字列、オプション): 会話/イベントが開始された時刻 (ISO 8601 形式)
finished_at （文字列、オプション）：会話/イベントが終了した時刻（ISO 8601形式）
language (文字列、オプション): 言語コード (デフォルト: "en")
geolocation （オブジェクト、オプション）: 会話の位置データ
latitude （数値）：緯度座標
longitude （数値）：経度座標
text_source_spec (文字列、オプション): ソースに関する追加の仕様
特定のユーザーのためにOmiに新しい思い出を作成します。
パラメータ:
user_id (文字列): 思い出を作成するユーザーID
text （文字列、オプション）：記憶を抽出するテキストコンテンツ
memories （配列、オプション）：直接作成される明示的なメモリオブジェクトの配列
content （文字列）：メモリの内容
tags (文字列の配列、オプション): メモリのタグ
text_source (文字列、オプション): テキストコンテンツのソース
text_source_spec (文字列、オプション): ソースに関する追加の仕様
MCP サーバーをテストするには、提供されているテスト クライアントを使用できます。
node test-mcp-client.jsこれにより、次の操作を実行できる対話型テスト クライアントが起動します。
会話を始める
思い出を手に入れる
会話を作成する
やめる
テスト クライアントは、すべての操作にデフォルトのテスト ユーザー ID ( test-user-123 ) を使用します。
MCPサーバーには、コンソールとログファイルの両方にログを書き込む組み込みのログ機能が搭載されています。これは、サーバーのアクティビティのデバッグや監視に役立ちます。
ログはプロジェクトディレクトリのlogs/mcp-server.logに書き込まれます。ログファイルには、タイムスタンプと以下の詳細情報が含まれます。
サーバーの起動とシャットダウン
すべてのAPIリクエストとレスポンス
エラーメッセージとスタックトレース
OmiへのAPI呼び出し
リクエストパラメータと応答データ
tailコマンドを使用して、ログをリアルタイムで表示できます。
tail -f logs/mcp-server.logこれにより、サーバーがリクエストを処理し、Omi API と対話するときにライブ更新が表示されます。
各ログエントリは次の形式に従います。
[2024-03-21T12:34:56.789Z] Log message hereタイムスタンプは ISO 8601 形式であるため、イベントの相関関係の特定や問題のデバッグが容易になります。
Creates a new Omi conversation with text content and metadata
| Name | Required | Description | Default |
|---|---|---|---|
| finished_at | No | When the conversation/event ended in ISO 8601 format. Optional. | |
| geolocation | No | Location data for the conversation. Optional object containing latitude and longitude. | |
| language | No | Language code (e.g., "en" for English). Optional, defaults to "en". | en |
| started_at | No | When the conversation/event started in ISO 8601 format. Optional. | |
| text | Yes | The full text content of the conversation | |
| text_source | Yes | Source of the text content. Required. Options: "audio_transcript", "message", "other_text". | |
| text_source_spec | No | Additional specification about the source. Optional. | |
| user_id | Yes | The user ID to create the conversation for |
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 states this is a creation operation, implying it's a write/mutation tool, but doesn't disclose any behavioral traits like permission requirements, rate limits, side effects, or what happens on success/failure. For a creation tool with zero annotation coverage, this leaves significant gaps in understanding how it behaves.
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 a single, efficient sentence that gets straight to the point without unnecessary words. It's appropriately sized for a creation tool, though it could potentially benefit from slightly more context given the lack of annotations and usage guidelines.
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 the complexity (8 parameters, nested objects, no output schema, and no annotations), the description is insufficient. It doesn't explain what a successful creation returns, error conditions, or behavioral constraints. For a creation tool with multiple parameters and no structured safety hints, more descriptive context is needed to be complete.
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?
The schema description coverage is 100%, so the schema already documents all 8 parameters thoroughly. The description adds minimal value beyond the schema by mentioning 'text content and metadata', which loosely maps to the 'text' and other fields. Since the schema does the heavy lifting, the baseline score of 3 is appropriate.
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 the verb 'creates' and the resource 'new Omi conversation', specifying it includes 'text content and metadata'. This distinguishes it from sibling tools like 'read_omi_conversations' (read vs. create) and 'create_omi_memories' (conversation vs. memories). However, it doesn't explicitly differentiate from 'create_omi_memories' beyond the resource name.
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 provides no guidance on when to use this tool versus alternatives. It doesn't mention when to choose this over 'create_omi_memories' or whether it's for initial conversation creation versus updates. There's no context about prerequisites, dependencies, or typical use cases.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Creates Omi memories by extracting from text or using explicit memory objects
| Name | Required | Description | Default |
|---|---|---|---|
| memories | No | An array of explicit memory objects to be created directly. Either this or text must be provided. | |
| text | No | The text content from which memories will be extracted. Either this or memories must be provided. | |
| text_source | No | Source of the text content. Optional. Options: "email", "social_post", "other". | |
| text_source_spec | No | Additional specification about the source. Optional. | |
| user_id | Yes | The user ID to create memories for |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden of behavioral disclosure. While it mentions the creation action and two input methods, it doesn't disclose important behavioral traits like whether this is a write operation (implied but not stated), what permissions are needed, whether it's idempotent, what happens on failure, or what the return format looks like. For a creation tool with zero annotation coverage, this leaves significant gaps in understanding how the tool behaves.
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 a single, efficient sentence that clearly states the tool's purpose and two key input methods. It's front-loaded with essential information and contains no redundant or unnecessary words, making it easy to parse quickly.
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 that this is a creation tool with no annotations and no output schema, the description is insufficiently complete. It doesn't explain what 'Omi memories' are in this context, what the tool returns (e.g., success/failure, created memory IDs), or any behavioral constraints (e.g., rate limits, authentication needs). For a tool with 5 parameters and significant functionality, more context is needed to use it effectively.
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%, meaning all parameters are well-documented in the input schema itself. The description adds minimal value beyond the schema by mentioning 'extracting from text' (hinting at the 'text' parameter) and 'using explicit memory objects' (hinting at the 'memories' parameter), but doesn't provide additional semantic context like examples, edge cases, or relationships between parameters. This meets the baseline for high schema coverage.
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 the action ('creates Omi memories') and specifies two methods ('extracting from text' or 'using explicit memory objects'), which gives a good sense of what the tool does. However, it doesn't differentiate itself from sibling tools like 'create_omi_conversation' or 'read_omi_memories', leaving some ambiguity about when to use this specific memory creation tool versus other memory/conversation tools.
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 mentioning two input methods ('extracting from text' or 'using explicit memory objects'), which provides some context for when to use it. However, it doesn't explicitly state when to choose this tool over alternatives like 'create_omi_conversation' or 'read_omi_memories', nor does it mention any prerequisites or exclusions. The guidance is present but incomplete.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Retrieves user conversations from Omi with pagination and filtering options
| Name | Required | Description | Default |
|---|---|---|---|
| include_discarded | No | Whether to include discarded conversations (default: false) | |
| limit | No | Maximum number of conversations to return (max: 1000, default: 100) | |
| offset | No | Number of conversations to skip for pagination (default: 0) | |
| statuses | No | Comma-separated list of statuses to filter conversations by | |
| user_id | Yes | The user ID to fetch conversations for |
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 mentions 'pagination and filtering options', which hints at some behavior, but fails to cover critical aspects like authentication requirements, rate limits, error handling, or what the return format looks like (e.g., JSON structure). For a retrieval tool with 5 parameters, this leaves significant gaps.
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 a single, efficient sentence that front-loads the core purpose ('retrieves user conversations from Omi') and adds relevant details ('with pagination and filtering options'). There is no wasted verbiage, making it highly concise and well-structured.
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 the complexity of a retrieval tool with 5 parameters, no annotations, and no output schema, the description is incomplete. It lacks details on authentication, rate limits, error cases, and the structure of returned data (e.g., conversation objects). Without annotations or an output schema, the agent has insufficient information to handle this tool effectively in context.
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%, so the input schema fully documents all 5 parameters with descriptions. The description adds minimal value by mentioning 'pagination and filtering options', which loosely corresponds to parameters like 'limit', 'offset', and 'statuses', but doesn't provide additional semantics beyond what the schema already specifies. This meets the baseline for high schema coverage.
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 the verb ('retrieves') and resource ('user conversations from Omi'), making the purpose evident. It also mentions 'pagination and filtering options' which adds specificity. However, it doesn't explicitly distinguish this tool from its sibling 'read_omi_memories', which might cause confusion about when to retrieve conversations versus 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?
The description provides no guidance on when to use this tool versus alternatives like 'read_omi_memories' or 'create_omi_conversation'. It mentions filtering options but doesn't specify scenarios or prerequisites for usage, leaving the agent without context for tool selection.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Retrieves user memories from Omi with pagination options
| Name | Required | Description | Default |
|---|---|---|---|
| limit | No | Maximum number of memories to return (max: 1000, default: 100) | |
| offset | No | Number of memories to skip for pagination (default: 0) | |
| user_id | Yes | The user ID to fetch memories for |
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 mentions 'pagination options', which adds some context about how results are handled, but it does not cover other aspects like rate limits, authentication needs, error conditions, or what the return format looks like. This leaves gaps in understanding the tool's behavior beyond basic retrieval.
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 a single, efficient sentence that directly states the tool's function and key feature (pagination). It is front-loaded with the core purpose and avoids unnecessary words, making it highly concise and well-structured for quick comprehension.
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 the tool's moderate complexity (3 parameters, no output schema, no annotations), the description is minimally adequate. It covers the basic purpose and hints at pagination but lacks details on return values, error handling, or usage context. This leaves the agent with incomplete information for effective tool invocation, though it meets a baseline for a read operation.
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?
The input schema has 100% description coverage, documenting all three parameters (limit, offset, user_id) with details like defaults and constraints. The description adds no additional meaning beyond this, as it only mentions 'pagination options' without elaborating on parameter usage. This meets the baseline for high schema coverage but does not enhance parameter understanding.
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 the verb ('Retrieves') and resource ('user memories from Omi'), making the purpose specific and understandable. However, it does not explicitly differentiate from sibling tools like 'read_omi_conversations', which might retrieve a different type of data, so it lacks sibling differentiation for a perfect score.
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 provides no guidance on when to use this tool versus alternatives, such as when to choose it over 'read_omi_conversations' or other siblings. It mentions pagination options but does not specify scenarios or prerequisites for usage, leaving the agent without contextual direction.
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.
v1.0.0create_omi_conversationcreate_omi_memoriesread_omi_conversationsread_omi_memoriesScored across 4 tools
Each tool has a clearly distinct purpose: create vs. read operations for two distinct resources (conversations and memories). The separation between conversations and memories is explicit, and the create/read distinction is unambiguous, leaving no room for confusion or misselection.
All tool names follow a consistent verb_noun pattern with 'create' or 'read' as the verb and 'omi_conversations' or 'omi_memories' as the noun. The naming is perfectly uniform, using snake_case throughout, making the set highly predictable and readable.
With 4 tools, the count is reasonable for a server focused on Omi conversations and memories. It covers create and read operations for both resources, which is well-scoped, though it might feel slightly thin if update or delete operations are expected in the domain, but it's not a significant issue.
The tool set provides create and read operations for both conversations and memories, covering basic CRUD elements. However, there are notable gaps: no update or delete tools for either resource, which could limit agent workflows if modifications or deletions are needed, making the surface incomplete for full lifecycle management.
A comprehensive Model Context Protocol (MCP) server that enables AI assistants to interact with yo…
A Model Context Protocol server for Wix AI tools
Persistent memory for AI agents to retain, retrieve, and recall conversation context through MCP.
An MCP memory server. One memory your agents share — across models, devices and apps.
