Open-source Indian mutual fund data tools trend now

What is driving the sudden interest

Open-source Indian mutual fund data tooling is getting attention across Reddit and developer circles. A big part of the buzz is that these tools focus on public, India-first sources like AMFI and mfapi.in. Several projects are positioned as “data engines” that can feed live market data directly into AI clients. The discussion is less about trading tips and more about plumbing - how to collect, validate, and serve financial data reliably. Another theme is convenience, with single-command installs and ready endpoints. Posts also highlight that many of these services do not require API keys. The momentum is amplified by the Model Context Protocol, which lets AI apps call tools in a standard way. Together, these ideas are pushing mutual-fund NAV collection from ad-hoc scripts into reusable infrastructure.

MCP servers: the new wrapper around market data

The Model Context Protocol, or MCP, is being used as a packaging layer for Indian finance data. In the shared context, MCP servers are described as tools that connect Claude Desktop and other MCP-compatible clients. The practical outcome is that an AI agent can query NAVs or market data without the user manually downloading files. This is being pitched as an alternative interface to classic APIs, not necessarily a replacement. Some projects run both an MCP endpoint and a normal REST API “mirror” side-by-side. This matters because not every consumer is an AI client, and HTTP integrations remain common. The same underlying data functions can be exposed to both MCP and REST consumers. The posts frame MCP as a way to make financial data callable, consistent, and agent-friendly. The trend is less about one vendor and more about standardization of access.

India Finance MCP: TypeScript server with a REST mirror

One trending project is “India Finance MCP”, positioned as an MCP server exposing Indian financial data to LLMs and agents. It is described as a single TypeScript codebase with two deployment targets. The MCP server uses FastMCP on Node and is deployed as an Apify Actor. In parallel, a REST mirror runs on Hono on Cloudflare Workers, aimed at standard HTTP clients. The author notes this REST surface can be monetized via RapidAPI, but the context focuses on availability rather than pricing. A highlighted capability is real-time mutual fund NAVs from AMFI, with a claim of 47,000+ schemes. The server also mentions GSTIN validation via checksum verification as an additional utility. Stock quotes are referenced as “coming in v1.1,” indicating the feature set is still evolving.

FinStack MCP: Python install, no key, many tools

Another widely shared tool is “FinStack MCP,” presented as an open-source financial data server for Indian markets. The project claims it connects Claude Desktop and any MCP client to live NSE, BSE, Nifty 50, Sensex, mutual fund NAV, FII/DII data, earnings calendar, and corporate actions. The creator frames it as a free alternative to expensive terminals, citing “Bloomberg charges $11,980/yr” as a comparison point in the post. It is described as free, with no API key required, and distributed via pip install finstack-mcp. The project also claims “58 tools for Indian markets,” including options Greeks and credit ratings. Compatibility is explicitly mentioned with Claude Desktop, Cursor, and Windsurf. The code is said to be live on GitHub, and the license is stated as MIT. Social posts also describe a simple setup flow that involves copying an MCP server JSON into a config.json.

Classic mutual fund APIs still matter: mfapi.in

Alongside MCP, developer conversations continue to cite mfapi.in as a foundational building block. mfapi.in is described as a free JSON API for India mutual fund data. The shared endpoints include https://www.mfapi.in/mf to list all schemes and https://www.mfapi.in/mf/{schemeCode} to fetch NAV history and metadata for a specific scheme. The context stresses that no authentication or API key is required. It also says data updates daily, making it suitable for routine refresh pipelines. A separate note describes an “India Mutual Fund MCP” powered by mfapi.in and exposed as an MCP service. That framing suggests mfapi.in is being wrapped for agent use, not replaced. The practical advantage is that many languages can consume it because it is plain HTTP returning JSON. In short, mfapi.in remains the simplest option when you want mutual fund history without adopting MCP.

Python libraries: mftool as a building block

For Python-first users, mftool is repeatedly referenced as an open-source library for Indian mutual fund data. It is described as fetching publicly available mutual-fund data, including NAV, scheme details, and returns. The feature list in the context includes last updated quotes by scheme code and outputs in DataFrame, JSON, and dictionary formats. It also mentions helper APIs to validate whether a scheme code is correct. Another emphasized capability is downloading all historical NAVs for a scheme. The library can list all schemes with their codes and compute daily scheme performance. The broader positioning is that it can support projects that need live quotes or large datasets for analytics. In the same cluster of repos, a portfolio management tool is also mentioned for tracking stocks, mutual funds, PPF, EPF, and other investments. The common thread is reuse: libraries reduce the need to scrape or parse raw AMFI files yourself.

Datasets: daily snapshots and large NAV histories

Beyond APIs, a dataset-style approach is trending for analysts who want bulk data rather than per-request lookups. One repository, “Mutual_Fund_Data,” is described as publishing daily-updated files fetched and processed via a Kaggle Notebook. It lists two primary artifacts: mutual_fund_data.csv and mutual_fund_nav_history.parquet. The CSV is described as a latest snapshot covering 9,000+ schemes, including NAV, AMC, AUM, and scheme categories, refreshed daily. The Parquet file is described as containing 20 million+ historical NAV records for time-series analysis. The context also claims 6,000+ new NAV records are added daily. Data sourcing is attributed to AMFI in the post. This dataset approach fits backtesting, comparisons, and research where repeated API calls are inefficient. It also creates a clearer separation between collection, storage, and downstream analysis.

Quick comparison table: what people are sharing

The current discussion highlights multiple ways to collect Indian mutual fund NAV data, each tuned for a different workflow. MCP servers are presented as ideal when the consumer is an AI agent or desktop AI client. REST APIs and public JSON endpoints fit web apps, scripts, and integrations. Python libraries target analysts and data engineers working locally with notebooks. Datasets target bulk analytics and historical research. The tools also vary by deployment style, from Apify Actors to Cloudflare Workers to pip packages. Some projects explicitly combine multiple interfaces to reduce lock-in. Below is a summary based only on the shared context.

Practical cautions and what to watch next

These tools are built on publicly available data, which is powerful but also requires careful handling. The context repeatedly emphasizes daily refresh cycles, which means “real-time” claims should be interpreted as the project describes them. Users should also note that different sources may represent scheme names, codes, and metadata slightly differently. In MCP setups, configuration is part of the workflow, and posts mention editing a config.json to add server JSON. REST mirrors simplify consumption but still require you to manage rate limits and stability, even when authentication is not required. The projects also show active development, such as India Finance MCP mentioning stock quotes in a future version. For FinStack MCP, the breadth of tools listed suggests a fast-moving scope, so users may want to pin versions in production. Dataset consumers should consider file sizes and update cadence when building pipelines. Overall, the trend is clear: Indian mutual fund data collection is shifting toward open, reusable components that can serve both apps and AI agents.