blocknative-mcp
# Blocknative MCP Server
An MCP server that provides real-time gas price predictions across multiple blockchains, powered by Blocknative.
## Features
- **Tools**:
- `predict_gas_price`: Fetches gas price predictions for a specified chain (default: Ethereum Mainnet, `chain_id=1`), including base fee and a table with confidence levels, price, max priority fee, and max fee.
- `estimate_gas_cost`: Estimates transaction costs based on gas limit, confidence level, and chain ID, returning costs in Gwei and ETH.
- `get_supported_chains`: Lists supported blockchains in a table with chain ID, system, and network.
- **Prompt**:
- `gas_price_query`: A prompt template for querying gas prices at a specific confidence level and chain ID.
- **Asynchronous**: Uses `httpx` for non-blocking HTTP requests to Blocknative's Gas Price and Chains APIs.
- **Optional API Key**: Supports low-frequency access without a Blocknative API key; high-frequency use requires setting `BLOCKNATIVE_API_KEY`.
## Prerequisites
- **Python**: Version 3.10
- **uv**: For dependency management and running the project ([installation guide](https://github.com/astral-sh/uv))
- **Blocknative API Key** (optional): Required for high-frequency API access. Sign up at [Blocknative](https://www.blocknative.com/) to obtain a free API key.
## Installation
1. **Clone the Repository**:
```bash
git clone https://github.com/kukapay/blocknative-mcp.git
cd blocknative-mcp
```
2. **Set Up the Project with `uv`**:
```bash
uv sync
```
3. **Set the Blocknative API Key (Optional)**:
- For high-frequency access, set the environment variable:
```bash
export BLOCKNATIVE_API_KEY="your-api-key-here"
```
- For low-frequency access, skip this step.
## Usage
The server provides three tools and one prompt, accessible via the FastMCP framework. You can run the server in development mode, execute it directly, or integrate it with Claude Desktop.
### Running the Server
1. **Development Mode with MCP Inspector**:
```bash
uv run mcp dev blocknative-mcp
```
This opens the MCP Inspector, a web interface for testing tools and prompts interactively.
2. **Direct Execution**:
```bash
uv run blocknative-mcp
```
3. **Claude Desktop Integration**:
```bash
uv run mcp install blocknative-mcp --name "Blocknative MCP Server"
```
This installs the server for use with Claude Desktop.
### Tools
- **`predict_gas_price(chain_id: int = 1) -> str`**:
Fetches gas price predictions for a chain (default: Ethereum Mainnet). Returns base fee and a Markdown table with confidence levels, price, max priority fee, and max fee.
- Example:
```markdown
Gas Price Predictions for Chain ID 1 (ethereum/main):
- Base Fee Per Gas: 0.382910791 Gwei
| Confidence | Price (Gwei) | Max Priority Fee (Gwei) | Max Fee (Gwei) |
|------------|--------------|-------------------------|----------------|
| 99% | 0.52 | 0.14 | 0.9 |
| 95% | 0.48 | 0.094 | 0.86 |
| 90% | 0.47 | 0.089 | 0.85 |
| 80% | 0.46 | 0.079 | 0.84 |
| 70% | 0.45 | 0.069 | 0.83 |
```
- **`estimate_gas_cost(gas_limit: int, confidence: int = 99, chain_id: int = 1) -> str`**:
Estimates transaction costs based on gas limit, confidence level, and chain ID. Returns costs in Gwei and ETH.
- Example:
```markdown
Estimated Gas Cost (Confidence 90%, Chain ID 1):
- Gas Limit: 21000
- Max Fee Per Gas: 0.85 Gwei
- Total Cost: 17850 Gwei (0.00001785 ETH)
```
- **`get_supported_chains(ctx: Optional[Context] = None) -> str`**:
Lists supported chains in a Markdown table with chain ID, system, and network.
- Example:
```markdown
Supported Chains:
| Chain ID | System | Network |
|------------|----------|-----------|
| 1 | ethereum | main |
| 137 | polygon | mainnet |
| 8453 | base | mainnet |
| 11155111 | ethereum | sepolia |
```
### Prompts
- **`gas_price_query(confidence: int, chain_id: int = 1) -> List[base.Message]`**:
A prompt template for querying gas prices at a specific confidence level and chain ID.
- Example:
```markdown
- User: What is the current gas price for chain ID 1 transactions with 90% confidence?
- Assistant: Let me fetch the gas price predictions for chain ID 1.
```
## License
This project is licensed under the MIT License. See the [LICENSE](LICENSE) file for details.
TDQS
Scored across 3 tools
Each tool has a clearly distinct purpose with no overlap: estimate_gas_cost calculates transaction costs, get_supported_chains lists available networks, and predict_gas_price forecasts prices. The descriptions clearly differentiate these gas platform functionalities, making tool selection unambiguous for an agent.
All three tools follow a consistent verb_noun naming pattern (estimate_gas_cost, get_supported_chains, predict_gas_price) with clear action-object relationships. The naming is uniform throughout, using snake_case consistently without any deviations or mixed conventions.
With only 3 tools, the server feels somewhat thin for a gas platform domain that could benefit from additional operations like historical gas price analysis, transaction simulation, or multi-chain batch operations. While the core functions are covered, the limited tool count may restrict agent capabilities in more complex scenarios.
The toolset covers the essential gas platform operations well: cost estimation, chain discovery, and price prediction. However, there are minor gaps such as the inability to simulate transactions with actual data, compare gas prices across multiple chains simultaneously, or access historical gas price trends, which could limit some advanced workflows.