AllRatesToday × DeepSeek — Currency Tools for DeepSeek Agents
Give DeepSeek chat and agents real-time access to currency exchange rates and historical data via function calling. Python package, one-line install.
allratestoday-deepseek is a Python package that plugs real-time currency exchange rates into DeepSeek chat and agents via function calling. DeepSeek's Chat Completions API is OpenAI-compatible, so the same tool-calling mechanism works — this package ships ready-to-use tool schemas, an agent wrapper, and a CLI.
Once installed, you can ask DeepSeek things like:
- "What's USD to EUR right now?"
- "Convert 2500 GBP to JPY."
- "How has USD/INR moved over the last 30 days?"
- "What's happening in FX markets today?"
allratestoday-deepseek on GitHub.
Works with deepseek-chat and deepseek-reasoner. Python 3.9+.
Install
pip install allratestoday-deepseek Quick start — one-shot
export DEEPSEEK_API_KEY=sk-xxxxx # from platform.deepseek.com/api_keys
allratestoday-deepseek --ask "What is 2500 USD in EUR right now?" Quick start — interactive
allratestoday-deepseek
# you: Convert 1000 GBP to JPY
# assistant: 1000 GBP ≈ 192,730 JPY at today's mid-market rate of 192.73. Library usage
from allratestoday_deepseek import DeepSeekCurrencyAgent
with DeepSeekCurrencyAgent() as agent:
print(agent.ask("How many Swiss Francs is 500 Japanese Yen?")) Low-level — drive the tool loop yourself
If you already have a DeepSeek conversation loop and just want the tools, import the schemas directly:
from openai import OpenAI
from allratestoday_deepseek import AllRatesTodayClient, TOOLS, dispatch_tool
llm = OpenAI(api_key=os.environ["DEEPSEEK_API_KEY"], base_url="https://api.deepseek.com")
rates = AllRatesTodayClient()
resp = llm.chat.completions.create(
model="deepseek-chat",
messages=[{"role": "user", "content": "Convert 100 USD to NGN"}],
tools=TOOLS,
tool_choice="auto",
)
for tc in resp.choices[0].message.tool_calls or []:
print(dispatch_tool(rates, tc.function.name, tc.function.arguments)) Tools exposed
| Tool | API key | What it does |
|---|---|---|
get_exchange_rate | no | Current mid-market rate between two currencies. |
convert_currency | no | Convert an amount between two currencies at the live rate. |
get_historical_rates | yes | Data over 1d / 7d / 30d / 1y. |
list_currencies | no | All supported currencies with codes, names, symbols. |
Environment variables
| Variable | Required? | Purpose |
|---|---|---|
DEEPSEEK_API_KEY | Yes | Your DeepSeek API key (platform.deepseek.com). |
ALLRATES_API_KEY | No | AllRatesToday key for higher limits. Get one at allratestoday.com/register. |
ALLRATES_BASE_URL | No | Override the API base URL. |
FAQ
Does this work with DeepSeek-R1?
Yes. Any DeepSeek model that supports function calling works — the package targets both
deepseek-chat and deepseek-reasoner.
Do I need an AllRatesToday API key?
Not for the public tools (get_exchange_rate, convert_currency, and
list_currencies). get_historical_rates and multi-target lookups
require an AllRatesToday key.
Why not just ask DeepSeek directly?
Without tool access, LLMs hallucinate rates — they don't know today's USD/EUR, only rough historical averages. This package ensures every rate comes from live data, not the model's guess.
Can I use it in a LangChain / LangGraph flow?
Yes — import TOOLS and dispatch_tool and wire them into whatever
orchestration framework you're using. The schemas are plain OpenAI-format JSON.
Install in 10 seconds
pip install, export keys, ask DeepSeek about real currency rates.
View on GitHub Get free AllRatesToday key