UluAlgorandMCP
Algorand ecosystem knowledge layer as an MCP server, enabling identification of applications and assets, protocol discovery, name resolution, and protocol summaries without on-chain queries.
README
UluAlgorandMCP
Algorand ecosystem MCP server. Returns meaning, not raw chain data.
What It Is
UluAlgorandMCP is an Algorand-specific knowledge and interpretation layer exposed as an MCP (Model Context Protocol) server. It provides protocol discovery, application identification, asset identification, naming resolution, and protocol summaries for the Algorand ecosystem.
What It Does
- Identifies known Algorand applications by app ID (protocol, role, purpose)
- Identifies known Algorand assets by asset ID (symbol, type, protocol association)
- Lists and describes Algorand protocols (DEXes, lending, naming, bridges, wallets)
- Resolves .algo names from a curated static registry
- Provides agent-friendly protocol summaries
What It Does Not Do
- Generic block/round lookup
- Account balance or state queries
- Transaction building, signing, or broadcasting
- Custody or key management
- Real-time on-chain data fetching
Those responsibilities belong to other layers:
UluCoreMCP → chain primitives (blocks, accounts, transactions)
UluAlgorandMCP → Algorand ecosystem meaning (this server)
UluWalletMCP → signing and custody
UluBroadcastMCP → transaction submission
Core returns facts. Algorand returns meaning.
How It Differs From UluCoreMCP
UluCoreMCP provides low-level chain primitives: look up a block, query an account, fetch a transaction. It is network-aware and returns raw chain data.
UluAlgorandMCP sits above that layer. It answers questions like "what is application 1002541853?" (Tinyman V2 Router) or "what protocols exist in the Algorand ecosystem?" without requiring any chain calls.
How It Mirrors UluVoiMCP
UluAlgorandMCP is the Algorand-specific sibling of UluVoiMCP. Both share:
- The same EmptyMCP scaffold
- The same project structure (
data/,lib/,tools/,index.js) - The same tool surface (10 tools across 3 modules)
- The same registry-backed architecture
- The same error conventions
UluVoiMCP covers the Voi ecosystem. UluAlgorandMCP covers Algorand.
Relationship to algorand-mcp
GoPlausible/algorand-mcp was used as a reference for understanding which Algorand ecosystem capabilities exist and which protocols are worth covering. However, UluAlgorandMCP does not replicate its architecture. algorand-mcp combines wallet, signing, submission, indexer access, and ecosystem integrations in one server. UluAlgorandMCP is deliberately smaller, focused only on the ecosystem knowledge layer.
Setup
npm install
Usage
node index.js
Adding to a Client
{
"mcpServers": {
"ulu-algorand-mcp": {
"command": "node",
"args": ["/absolute/path/to/UluAlgorandMCP/index.js"]
}
}
}
Tools
Protocol Discovery
| Tool | Description |
|---|---|
get_protocols |
List all known Algorand protocols. Optional type filter. |
get_protocol |
Get full metadata for a protocol by ID. |
get_protocol_contracts |
List known contracts and assets for a protocol. |
get_protocol_summary |
Get a concise agent-friendly protocol summary. |
Application & Asset Identification
| Tool | Description |
|---|---|
identify_application |
Identify an Algorand app by ID — returns protocol, role, type. |
identify_asset |
Identify an Algorand asset by ID — returns symbol, type, tags. |
get_contract_role |
Get the known role of an Algorand application. |
Naming Resolution
| Tool | Description |
|---|---|
resolve_name |
Resolve a .algo name from the static registry. |
reverse_resolve_address |
Look up names associated with an address. |
search_names |
Search the name registry by substring pattern. |
Example Requests and Responses
List all DEX protocols
Request:
{ "type": "dex" }
Response:
{
"protocols": [
{
"id": "tinyman",
"name": "Tinyman",
"type": "dex",
"description": "Leading Algorand DEX using constant-product AMM pools...",
"tags": ["defi", "amm", "swap", "liquidity"]
}
]
}
Identify an application
Request:
{ "appId": 1002541853 }
Response:
{
"appId": 1002541853,
"recognized": true,
"name": "Tinyman V2 Router",
"protocol": "tinyman",
"protocolName": "Tinyman",
"role": "amm-router",
"type": "dex",
"description": "Tinyman V2 AMM router and validator..."
}
Identify an asset
Request:
{ "assetId": 31566704 }
Response:
{
"assetId": 31566704,
"recognized": true,
"name": "USDC",
"symbol": "USDC",
"decimals": 6,
"type": "stablecoin",
"protocol": null,
"protocolName": null,
"tags": ["stablecoin", "circle", "usd"],
"description": "USD Coin issued by Circle. The primary USD stablecoin on Algorand."
}
Resolve a name
Request:
{ "name": "tinyman.algo" }
Response:
{
"name": "tinyman.algo",
"description": "Tinyman — leading Algorand DEX",
"source": "static-registry"
}
Identify an unknown application
Request:
{ "appId": 999999999 }
Response:
{
"appId": 999999999,
"recognized": false,
"message": "Application 999999999 is not in the known Algorand registry."
}
Project Structure
index.js Server entry point
package.json Dependencies and metadata
data/
protocols.json Curated protocol registry
applications.json Known application IDs and roles
assets.json Known asset IDs and metadata
names.json Well-known .algo names
lib/
errors.js Tool result/error helpers
registry.js Data loading and lookup functions
tools/
protocols.js get_protocols, get_protocol, get_protocol_contracts, get_protocol_summary
identify.js identify_application, identify_asset, get_contract_role
names.js resolve_name, reverse_resolve_address, search_names
Initial Registry Coverage
The v1 registry includes curated entries for:
- DEX/AMM: Tinyman, Pact, HumbleSwap, CompX
- Lending: Folks Finance
- Naming: NFDomains
- Analytics: Vestige
- Bridge: Algomint, Aramid Bridge
- Liquid Staking: Cometa
- Real World Assets: Lofty
- Oracle: Goracle
- Wallets: Pera, Defly
The registry is static and curated. Dynamic enrichment can be added in future versions.
Constraints
- JavaScript only, no TypeScript
- No bundlers or build systems
- Lightweight, stdio MCP server
- Extended from the EmptyMCP scaffold
- Structurally consistent with all Ulu MCP servers
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。