Skip to content

Messari action provider #549

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 5 additions & 0 deletions typescript/.changeset/ninety-gifts-know.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,5 @@
---
"@coinbase/agentkit": patch
---

Add a new Messari action provider that enables AI agents to query the Messari AI toolkit for crypto market research data.
1 change: 1 addition & 0 deletions typescript/agentkit/src/action-providers/index.ts
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@ export * from "./erc20";
export * from "./erc721";
export * from "./farcaster";
export * from "./jupiter";
export * from "./messari";
export * from "./pyth";
export * from "./moonwell";
export * from "./morpho";
Expand Down
103 changes: 103 additions & 0 deletions typescript/agentkit/src/action-providers/messari/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,103 @@
# Messari Action Provider

The Messari Action Provider enables AI agents to query the [Messari AI toolkit](https://messari.io/) for crypto market research data. This provider allows agents to ask research questions about market data, statistics, rankings, historical trends, and information about specific protocols, tokens, or platforms.

## Getting an API Key

To use the Messari Action Provider, you need to obtain a Messari API key by following these steps:

1. Sign up for a Messari account at [messari.io](https://messari.io/)
2. After signing up, navigate to [messari.io/account/api](https://messari.io/account/api)
3. Generate your API key from the account dashboard

For more detailed information about authentication, refer to the [Messari API Authentication documentation](https://docs.messari.io/reference/authentication).

Different subscription tiers provide different levels of access to the API. See the [Rate Limiting](#rate-limiting) section for details.

## Configuration

Once you have your Messari API key, you can configure the provider in two ways:

### 1. Environment Variable

Set the `MESSARI_API_KEY` environment variable:

```bash
MESSARI_API_KEY=your_messari_api_key
```

### 2. Direct Configuration

Pass the API key directly when initializing the provider:

```typescript
import { messariActionProvider } from "@coinbase/agentkit";

const provider = messariActionProvider({
apiKey: "your_messari_api_key",
});
```

## Rate Limiting

The Messari API has rate limits based on your subscription tier:

| Subscription Tier | Daily Request Limit |
|-------------------|---------------------|
| Free (Unpaid) | 2 requests per day |
| Lite | 10 requests per day |
| Pro | 20 requests per day |
| Enterprise | 50 requests per day |

If you need more than 50 requests per day, you can contact Messari's sales team to discuss a custom credit allocation system for your specific needs.

## Actions

### `research_question`

This action allows the agent to query the Messari AI toolkit with a research question about crypto markets, protocols, or tokens.

#### Input Schema

| Parameter | Type | Description |
|-----------|--------|--------------------------------------------------------------|
| question | string | The research question about crypto markets, protocols, or tokens |

#### Example Usage

```typescript
import { AgentKit, messariActionProvider } from "@coinbase/agentkit";
import { getLangChainTools } from "@coinbase/agentkit-langchain";
import { createReactAgent } from "@langchain/langgraph/prebuilt";
import { ChatOpenAI } from "@langchain/openai";

// Initialize AgentKit with the Messari action provider
const agentkit = await AgentKit.from({
actionProviders: [messariActionProvider()],
});

// Get LangChain tools from AgentKit
const tools = await getLangChainTools(agentkit);

// Create a LangChain agent with the tools
const llm = new ChatOpenAI({ model: "gpt-4o-mini" });
const agent = createReactAgent({
llm,
tools,
});

// The agent can now use the Messari research_question action
// Example prompt: "What is the current price of Ethereum?"
```

#### Example Response

```
Messari Research Results:

Ethereum (ETH) has shown strong performance over the past month with a 15% price increase. The current price is approximately $3,500, up from $3,000 at the beginning of the month. Trading volume has also increased by 20% in the same period.
```

## Network Support

The Messari Action Provider is network-agnostic, meaning it supports all networks. The research capabilities are not tied to any specific blockchain network.
2 changes: 2 additions & 0 deletions typescript/agentkit/src/action-providers/messari/index.ts
Original file line number Diff line number Diff line change
@@ -0,0 +1,2 @@
export * from "./messariActionProvider";
export * from "./schemas";
Original file line number Diff line number Diff line change
@@ -0,0 +1,200 @@
import { messariActionProvider, MessariActionProvider } from "./messariActionProvider";

const MOCK_API_KEY = "messari-test-key";

// Sample response for the research question action
const MOCK_RESEARCH_RESPONSE = {
data: {
messages: [
{
role: "assistant",
content:
"Ethereum (ETH) has shown strong performance over the past month with a 15% price increase. The current price is approximately $3,500, up from $3,000 at the beginning of the month. Trading volume has also increased by 20% in the same period.",
},
],
},
};

// Sample error response in Messari format
const MOCK_ERROR_RESPONSE = {
error: "Internal server error, please try again. If the problem persists, please contact support",
data: null,
};

describe("MessariActionProvider", () => {
let provider: MessariActionProvider;

beforeEach(() => {
process.env.MESSARI_API_KEY = MOCK_API_KEY;
provider = messariActionProvider({ apiKey: MOCK_API_KEY });
jest.restoreAllMocks();
});

afterEach(() => {
jest.clearAllMocks();
delete process.env.MESSARI_API_KEY;
});

describe("constructor", () => {
it("should initialize with API key from constructor", () => {
const customProvider = messariActionProvider({ apiKey: "custom-key" });
expect(customProvider["apiKey"]).toBe("custom-key");
});

it("should initialize with API key from environment variable", () => {
process.env.MESSARI_API_KEY = "env-key";
const envProvider = messariActionProvider();
expect(envProvider["apiKey"]).toBe("env-key");
});

it("should throw error if API key is not provided", () => {
delete process.env.MESSARI_API_KEY;
expect(() => messariActionProvider()).toThrow("MESSARI_API_KEY is not configured.");
});
});

describe("researchQuestion", () => {
it("should successfully fetch research results", async () => {
const fetchMock = jest.spyOn(global, "fetch").mockResolvedValue({
ok: true,
json: async () => MOCK_RESEARCH_RESPONSE,
} as Response);

const question = "What is the current price of Ethereum?";
const response = await provider.researchQuestion({ question });

// Verify the API was called with the correct parameters
expect(fetchMock).toHaveBeenCalled();
const [url, options] = fetchMock.mock.calls[0];

// Check URL
expect(url).toBe("https://api.messari.io/ai/v1/chat/completions");

// Check request options
expect(options).toEqual(
expect.objectContaining({
method: "POST",
headers: expect.objectContaining({
"Content-Type": "application/json",
"x-messari-api-key": MOCK_API_KEY,
}),
body: JSON.stringify({
messages: [
{
role: "user",
content: question,
},
],
}),
}),
);

// Check response formatting
expect(response).toContain("Messari Research Results:");
expect(response).toContain(MOCK_RESEARCH_RESPONSE.data.messages[0].content);
});

it("should handle non-ok response with structured error format", async () => {
const errorResponseText = JSON.stringify(MOCK_ERROR_RESPONSE);

jest.spyOn(global, "fetch").mockResolvedValue({
ok: false,
status: 500,
statusText: "Internal Server Error",
text: async () => errorResponseText,
} as Response);

const response = await provider.researchQuestion({
question: "What is the current price of Bitcoin?",
});

// Should use the structured error message from the response
expect(response).toContain("Messari API Error: Internal server error");
expect(response).not.toContain("500"); // Should not include technical details when we have a structured error
});

it("should handle non-ok response with non-JSON error format", async () => {
const plainTextError = "Rate limit exceeded";

jest.spyOn(global, "fetch").mockResolvedValue({
ok: false,
status: 429,
statusText: "Too Many Requests",
text: async () => plainTextError,
} as Response);

const response = await provider.researchQuestion({
question: "What is the current price of Bitcoin?",
});

// Should fall back to detailed error format
expect(response).toContain("Messari API Error:");
expect(response).toContain("429");
expect(response).toContain("Too Many Requests");
expect(response).toContain(plainTextError);
});

it("should handle JSON parsing error in successful response", async () => {
jest.spyOn(global, "fetch").mockResolvedValue({
ok: true,
json: async () => {
throw new Error("Invalid JSON");
},
} as unknown as Response);

const response = await provider.researchQuestion({
question: "What is the market cap of Solana?",
});

expect(response).toContain("Unexpected error: Failed to parse API response");
expect(response).toContain("Invalid JSON");
});

it("should handle invalid response format", async () => {
jest.spyOn(global, "fetch").mockResolvedValue({
ok: true,
json: async () => ({ data: { messages: [] } }), // Empty messages array
} as Response);

const response = await provider.researchQuestion({
question: "What is the market cap of Solana?",
});

expect(response).toContain(
"Unexpected error: Received invalid response format from Messari API",
);
});

it("should handle fetch error", async () => {
const error = new Error("Network error");
jest.spyOn(global, "fetch").mockRejectedValue(error);

const response = await provider.researchQuestion({
question: "What is the market cap of Solana?",
});

expect(response).toContain("Unexpected error: Network error");
});

it("should handle string error with JSON content", async () => {
// This simulates a case where an error might be stringified JSON
const stringifiedError = JSON.stringify(MOCK_ERROR_RESPONSE);
jest.spyOn(global, "fetch").mockRejectedValue(stringifiedError);

const response = await provider.researchQuestion({
question: "What is the market cap of Solana?",
});

// Should parse the JSON string and extract the error message
expect(response).toContain("Messari API Error: Internal server error");
});
});

describe("supportsNetwork", () => {
it("should always return true as research is network-agnostic", () => {
expect(provider.supportsNetwork({ protocolFamily: "evm" })).toBe(true);
expect(provider.supportsNetwork({ protocolFamily: "solana" })).toBe(true);
expect(provider.supportsNetwork({ protocolFamily: "unknown" })).toBe(true);
});
});
});
Loading