Strategy Report
The `buy_then_extend` path perfectly aligns with the strategy of treating market data as an enabling capability rather than a core differentiator. By licensing raw data feeds from authorized commercial vendors via standard APIs and building the analytics and compliance workflows internally over TimescaleDB, the platform avoids massive infrastructure overhead while maintaining precise control over corporate action accuracy and SEBI compliance.
Capability
Market data infrastructure layer providing historical OHLCV data (NSE and BSE equities, daily and 1-minute bars for at least 20 years), machine-readable corporate actions, index and custom index composition history, 15-minute delayed or sub-second real-time tick feeds, AMFI mutual fund NAVs, and F&O market data.
Context
Series A fintech startup building an institutional-grade portfolio analytics and compliance platform for Indian wealth managers (boutique advisory firms, family offices, and SEBI-registered RIAs managing ₹50 crore to ₹2,000 crore in client AUM).
Problem
Calculating accurate historical performance, trailing returns, and backtesting requires point-in-time adjusted historical prices accounting perfectly for corporate actions. Incorrect calculations lead to client-facing legal liabilities. Furthermore, commercial redistribution of exchange data requires strict, authorized exchange licensing, which limits self-built solutions due to legal and compliance barriers.
Decisive Factors
Infrastructure Cost & Complexity
Building direct exchange connectivity costs over 4 lakhs/month for co-location and 38.5 lakhs in fixed fees, making buy_then_extend far more viable.
Vendor Licensing Constraints
Restrictive licenses from vendors like TrueData block commercial redistribution, necessitating careful vendor selection (e.g., Accord Fintech NXT) to power the internal extension.
Capability Intent Focus
The platform's core moat is its analytics and compliance workflow; licensing the raw market data enables engineering to focus entirely on building that differentiating layer.
Architecture & Portfolio Synergies
Extending a bought feed via internal TimescaleDB/Kafka ingestion allows the resulting market-event stream to natively feed the parallel vector-database project without creating redundant ingestion paths.
Path Analysis
Pros
Cons
- ✕Must aggressively filter out vendors like TrueData due to restrictive single-subscriber licenses that explicitly block institutional platform redistribution.c4225f
- ✕National Stock Exchange binary snapshot data requires complex custom parsing over SFTP if used to augment vendor feeds directly.4da0b4ccabdf
Key Risks
Reversibility
↩ Highly reversible; integrating via standard API/FTP protocols from vendors like Accord Fintech NXT ensures the architecture avoids lock-in if price hikes exceed the ₹8-10 lakh/month ceiling.6a4b6c
Second Opinion
Wins when:
- A commercial vendor emerges with a fully managed, compliant, and correctly adjusted historical database that natively supports institutional backtesting and redistribution, removing the need for any internal pipeline extensions.
Open Questions
- Are Accord Fintech NXT's historical OHLCV and corporate action datasets perfectly adjusted point-in-time for algorithmic backtesting?
- What are the exact latency SLAs for Accord Fintech NXT's API/FTP feeds, and do they meet the sub-second tick requirement for alerts?
- Does the mfapi.in platform have the necessary reliability and legal backing for institutional-grade platform use, or is a commercial mutual fund data vendor required?
- How does the parallel vector-database project expect the upstream market-event stream to be specifically structured for its semantic alerts?
Evidence Index
| ID | Claim / Source | Status | Track | Dim | Flags |
|---|---|---|---|---|---|
054fae | Agreements for sharing real-time price data with third parties must be reviewed by the MII or intermediary Board annually. “reviewed by the Board of the MIIs or market intermediaries at least once” New SEBI Norms for sharing of real time price data to third parties 2024-05-24 | PARTIAL | BUILD | m9 | partial_evidence |
0cdf28 | Setting up a self-hosted production ClickHouse cluster requires managing configuration of replication, sharding, and ZooKeeper coordination. “Setting up a production-ready ClickHouse® cluster involves configuring replication, sharding, and ZooKeeper coordination” Honest guide to the best ClickHouse® alternatives in 2026 2026-05-18 | SUPPORTED | BUILD | m4 | |
4a4761 | SEBI mandates that no real-time price data be shared with third parties except for orderly market functioning or regulatory requirements. “no real time price data is shared with any third party” New SEBI Norms for sharing of real time price data to third parties 2024-05-24 | PARTIAL | BUILD | m9 | partial_evidence |
83ac9f | The National Stock Exchange provides real-time data online through dedicated leased line circuits owned by the customer. “It is provided on-line through a dedicated leased line circuit.” 2026-06-16 | SUPPORTED | BUILD | m14 | |
928202 | Point-in-time corporate actions data enables developers to backtest algorithms accurately without survivorship or lookahead bias. “Backtest accurately without survivorship or lookahead bias.” Corporate Actions Data & APIs | Databento 2022-06-05 | SUPPORTED | BUILD | m1 | |
ac2118 | NSE offers 15-minute delayed snapshot data files generated on a regular 1-minute interval basis. “generated with 15 minutes delay on regular 1 minute interval basis” 2026-06-16 | SUPPORTED | BUILD | m14 | |
e03a14 | Delayed or incorrect data in data pipelines can manifest as expensive and time-consuming user-facing issues. “Delayed or incorrect data in your pipeline can manifest in user-facing issues” 2018-01-01 | SUPPORTED | BUILD | m4 | stale_cost |
ef33d4 | TimescaleDB provides PostgreSQL users with automatic partitioning and full SQL support for time-series analytics. “Full SQL support, automatic partitioning” Honest guide to the best ClickHouse® alternatives in 2026 2026-05-18 | SUPPORTED | BUILD | m13 | |
fc6ff2 | ClickHouse is not fully ACID compliant, which is an important operational limitation to consider for transaction-heavy workloads. “ClickHouse does not offer full ACID compliance” ClickHouse Alternatives 2026: Real-Time OLAP DBs 2025-01-09 | PARTIAL | BUILD | m13 | partial_evidence |
02c3e3 | To distribute tick-by-tick market data, vendors must pay a fixed fee of 38.5 lakhs to the National Stock Exchange. “paid the fixed fees of 38.5lks and spent on infra.” How do I get tick by tick data directly via NSE 2019-02-20 | PARTIAL | BUY | m5 | partial_evidence, stale_cost |
3105d7 | Deploying a single server rack in the exchange co-location for direct tick-by-tick data access can cost over 4 lakhs per month. “cost you upwards of 4lks a month.” How do I get tick by tick data directly via NSE 2019-02-20 | PARTIAL | BUY | m5 | partial_evidence, stale_cost |
38512a | TrueData delivers market data APIs supporting NSE EQ, BSE EQ, and derivative segments. “delivering services for NSE EQ, NSE Indices, NSE F&O, BSE EQ, BSE Indices, BSE F&O” Real-Time Market Data API for NSE, BSE & MCX | Low Latency APIs 2025-01-01 | SUPPORTED | BUY | m10 | |
47d388 | TrueData maintains a strict non-refundable payment policy once an order is placed or data is delivered. “Amount paid will not be refunded once order is placed / data is delivered.” Real-Time Market Data API for NSE, BSE & MCX | Low Latency APIs 2025-01-01 | SUPPORTED | BUY | m5 | stale_cost |
4da0b4 | National Stock Exchange snapshot data files are provided in binary format over the internet. “Snapshot Data is provided in binary file format over the internet” 2026-06-16 | SUPPORTED | BUY | m14 | |
586f64 | The mfapi.in platform updates its historical mutual fund NAV data six times per day. “Updated 6x daily (10:05 AM, 2:05 PM, 6:05 PM, 9:05 PM, 3:09 AM, 5:05 AM IST)” 2025-01-01 | PARTIAL | BUY | m10 | partial_evidence |
65f322 | Streaming live tick-by-tick market data over the internet is constrained by large file sizes reaching tens of gigabytes. “It is tens of GB’s worth of data” How do I get tick by tick data directly via NSE 2019-02-20 | SUPPORTED | BUY | m14 | |
6a4b6c | Accord Fintech NXT provides stock market and mutual fund information feeds through API and FTP protocols. “We provide financial information feed through FTP and API.” | PARTIAL | BUY | m14 | partial_evidence |
a000ed | TrueData does not permit using its market data for gaming or virtual trading/simulation applications without exchange approval. “TrueData does not provide market data for any kind of Gaming / Virtual trading” Real-Time Market Data API for NSE, BSE & MCX | Low Latency APIs 2025-01-01 | SUPPORTED | BUY | m9 | |
a173a2 | The mfapi.in platform offers unrestricted access to Indian mutual fund data without requiring API keys or rate limiting. “No authentication, no API keys, no rate limiting. Just pure data access.” 2025-01-01 | PARTIAL | BUY | m10 | partial_evidence |
c4225f | TrueData restricts the license of its market data feed strictly to single-subscriber personal charting use. “This Data is for Personal Charting Use of a Single Subscriber only.” Real-Time Market Data API for NSE, BSE & MCX | Low Latency APIs 2025-01-01 | SUPPORTED | BUY | m9 | |
ccabdf | Downloading snapshot files from the National Stock Exchange requires the use of SFTP protocol. “requires the use of an SFTP protocol to download these files.” 2026-06-16 | SUPPORTED | BUY | m14 | |
e6e417 | TrueData offers market data API ingestion support using both WebSockets and RESTful protocols. “WebSockets, RESTful, DotNet, COM” Real-Time Market Data API for NSE, BSE & MCX | Low Latency APIs 2025-01-01 | PARTIAL | BUY | m14 | partial_evidence |