kFinance

kFinance

The kFinance Python library provides a simple interface for the LLM-ready API, streamlining API requests and response handling. It can be used on its own, with LLMs, or integrated into applications.

For a complete overview of the functions, usage, and features of the kFinance Python library, please refer to documentation here.

Any questions or suggestions can be sent to the kFinance Maintainers.

Setup

You can install kFinance on PyPI via

pip install kensho-kfinance

Getting started

To receive access, please email S&P Global Market Intelligence for information on free trials and pricing.

Once access is obtained, get started using the Authentication Guide and Usage Guide.

To get started, we provide some notebooks:

• The LLM-ready API Basic Usage notebook demonstrates how fetch data with the kFinance client.
• The tool_calling notebooks show how the kFinance library can be used for tool calling. We provide notebooks for OpenAI (GPT), Anthropic (Claude), and Google (Gemini). Each of these integrations comes in a langchain version, which uses langchain as a wrapper to simplify the integration, and as a lower level non-langchain version.

We also provide an interactive notebook that demonstrates some usage examples.

MCP (Model Context Protocol)

To run the kFinance MCP server use:

python -m kfinance.mcp

This function initializes and starts an MCP server that exposes the kFinance tools. The server supports multiple authentication methods and transport protocols to accommodate different deployment scenarios.

The server's full signature is as follows:

kfinance.mcp [--stdio|-s|--sse|--streamable-http] --refresh-token <refresh-token> --client-id <client-id> --private-key <private-key> --kid <kid> --api-host <api-host>

Authentication Methods (in order of precedence):

1. Refresh Token: Uses an existing refresh token for authentication. The --refresh-token <refresh-token> argument must be provided.
2. Key Pair: Uses client ID and private key for authentication. Both the --client-id <client-id> and --private-key <private-key> arguments must be provided. Add --kid <kid> when more than one public key is registered, since Okta rejects an assertion without a key ID in that case.
3. Browser: Falls back to browser-based authentication flow. This occurs if no auth arguments are provided.

Transport Layers:

• --stdio / -s: Standard input/output transport (use with MCP Inspector)
• --sse: Server-Sent Events transport (default)
• --streamable-http: HTTP transport

API Host:

• --api-host <api-host>: The kFinance API host to fetch data from. Defaults to production (https://kfinance.kensho.com). Useful for pointing the server at a local or pre-production backend.
• Refresh tokens are redeemed against --api-host, so a token issued by one host will not authenticate against another. Key-pair authentication goes to Okta instead, and so works against any host.

Examples:

# Using stdio with MCP Inspector
npx @modelcontextprotocol/inspector python -m kfinance.mcp --stdio --refresh-token <token>

# Using SSE (default)
python -m kfinance.mcp --refresh-token <token>

# Using streamable-http
python -m kfinance.mcp --streamable-http --refresh-token <token>

# Against a locally running backend
python -m kfinance.mcp --stdio --api-host http://localhost:8000 --refresh-token <token>

MCP Proxy

The kFinance MCP proxy is a skeleton server that forwards requests to a remote kfinance MCP backend, injecting authentication tokens into every outgoing request. It is intended as a starting point for building a full production MCP proxy service.

# Using a refresh token (for experimentation)
AUTH_REFRESH_TOKEN=<token> python -m kfinance.proxy_mcp

# Using a key pair (for production)
AUTH_CLIENT_ID=<client-id> AUTH_PRIVATE_KEY=<private-key> python -m kfinance.proxy_mcp

# Add AUTH_KID when more than one public key is registered
AUTH_CLIENT_ID=<client-id> AUTH_PRIVATE_KEY=<private-key> AUTH_KID=<kid> python -m kfinance.proxy_mcp

The server starts on http://127.0.0.1:8000/mcp by default. Use --host and --port to configure binding.

For full documentation including configuration options, authentication methods, and production considerations, see kfinance/integrations/proxy_mcp/README.md.

Development

Working with Local Package Version

If you need to develop using a local version of the kFinance package in another project that uses poetry for package and dependency management, follow these steps:

1. In your dependent project's pyproject.toml, replace the kFinance package version specification:

   # Replace this:
   # kensho-kfinance = "~2.0.1"
   # With this:
   kensho-kfinance = { path = "/absolute/path/to/kfinance", develop = true }
   

   The develop = true flag ensures that the package always matches your local version without requiring reinstallation after changes.

2. Update your project's dependencies:

   poetry update    # Update poetry.lock with new changes
   poetry install   # Install dependencies with updated configuration
   

   If you encounter the error "your pyproject.toml file has significantly changed", run:

   poetry lock     # Sync the changes to poetry.lock
   

Versioning

The kFinance uses semantic versioning (major, minor, patch). To bump the version, add a new entry in CHANGELOG.md. This will generate a new version of the library as part of the release process.

License

Licensed under the Apache 2.0 License. Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Copyright 2025-present Kensho Technologies, LLC. The present date is determined by the timestamp of the most recent commit in the repository.