This agent provides real-time pricing and other cryptoasset metrics by pulling data from CoinGecko and DefiLlama APIs.
It currently supports the following metrics:
- current price of coins
- current price of NFT collections
- market cap
- fully diluted valuation
- total value locked
It is possible to ask questions about assets by referring to them either by their name or their ticker symbol.
This is the recommended way to run the project in order to avoid dependency hell. You will need to build then run:
cd agent
docker build -t agent .
If you are using Apple Silicon then you may experience problems due to the base image not being build for arm64. We have included a separate Dockerfile in order to deal with this issue, run:
docker build . -t agent -f Dockerfile-apple
To run:
docker run --name agent -p 5000:5000 agent
cd agent
pip install -r requirements.txt
pip install --no-cache-dir -U llama-cpp-python
Docker exposes an API on port 5000 to communicate with the agent.
url = 'http://127.0.0.1:5000/'
data = {'prompt': "what is the market cap of dogwifhat"}
response = requests.post(url, json=data)
print(response.text)
A notebook has been provided to run test queries against the API:
Test Docker.ipynb
To use the API run
python agent.py
To use an interactive CLI prompt to test the agent run:
python agent-cli.py
-
what is the price of etherum
-
price of eth
-
what is the market cap of dogwifhat
-
mc of wif
-
what is the fully diluted valuation of solana
-
fdv sol
-
what is the total value locked in sushi
-
tvl of pendle
This project uses the openhermes-2.5-mistral-7b GGUF model and performs well on a modern laptop using only CPU.
The CoinGecko search API is used to find the asset that is being referenced. In case multiple matching assets are found, the agent will select the one with the largest market cap
When consuming this API as part of a larger agent, care should be taken to ensure that responses do not pass thorugh an LLM that hallucinates the number before the response is sent to the user.