# pandas-ta The DataFrame indicator library whose repo is gone and whose last release is a year old. *https://stockmarketstack.com/tools/pandas-ta · Market Analysis & Portfolio Optimization Libraries* ## Facts ### At a glance | Field | Value | | --- | --- | | Vendor | Kevin Johnson | | Category | Market Analysis & Portfolio Optimization Libraries | | Job | indicators | | Website | https://pypi.org/project/pandas-ta/ | | Pricing model | open-source | | Free tier | true | | Open source | true | | Licence | MIT | | Self-hosted | true | | Tested hands-on | false | | Last updated | 2026-09-19 | ### Coverage | Field | Value | | --- | --- | | Asset classes | stocks, etf, futures, forex, crypto, indices | | Markets | global | | Works outside the US | true | | Data latency | none | | Platforms | library | | AI features | none | ### Interfaces | Field | Value | | --- | --- | | API | false | | Webhooks | false | | Scripting | Python | | Python | true | | Spreadsheet add-in | false | | MCP server | false | | Export | none | ### Capabilities Yes: none No: charting, screening, scanning, backtesting, automation, live_trading, paper_trading, portfolio_tracking, broker_import, tax_reporting, alerts, news, options_analysis *Verified: pricing 2026-09-19; capabilities 2026-09-19; coverage 2026-09-19.* ## What it is pandas-ta registers a `.ta` accessor on any pandas DataFrame, so a column of indicators is one call — `df.ta.macd()`, `df.ta.bbands()`, `df.ta.strategy("all")` — with no binary dependency and no array wrangling. That ergonomic beat every alternative for years, and it is why the library is wired into so much existing code. It is also, as of September 2026, a package with no project behind it. The last release on PyPI is **0.4.71b0, uploaded 14 September 2025** — a year old, and still carrying a beta version string seven years in. The GitHub repository at `twopirllc/pandas-ta` **returns 404**; the Internet Archive last captured it working on 2 April 2025 and captured a 404 on 25 June 2025. The account itself still exists and now shows zero public repositories, so the issue tracker, the history and the fork point are all gone with it. The documentation site at `pandas-ta.dev` resolves in DNS but does not answer, which means the links in the package's own README go nowhere. PyPI lists one maintainer account, `pta`. PyPI also carries exactly two versions of the package now — `0.4.67b0` and `0.4.71b0`, both from September 2025. Every earlier release has been deleted from the index, including the `0.3.14b0` that years of requirements files pin, so an existing pinned install no longer resolves at all. The archived home page explains the direction of travel. In its last working capture, on 11 October 2025, the project wrote that funding levels were "unsustainable and risks discontinuation", and that low funding had reduced the library to "a yearly public release with maintenance updates", paid for by donations, sponsorships and yearly subscriptions. No yearly release has appeared since. ## Pricing The artefact on PyPI is free and MIT-licensed — the LICENSE file inside the 0.4.71b0 sdist reads "Copyright (c) 2019+ Kevin Johnson", although the PyPI metadata carries no licence field or classifier at all. There is nothing on sale that can be reached: the subscription the project described is announced only on a site that no longer responds. ## Data & coverage No data. The library operates on a DataFrame you supply, indexed by date with OHLCV columns, and never touches a market or a ticker. TA-Lib, if installed, is used for the candlestick patterns and as an optional faster path for some core indicators. ## Limitations - `numba==0.61.2` is pinned exactly rather than bounded. That release shipped wheels for CPython 3.10–3.13 only, and the metadata requires 3.12 or newer, so the usable window is Python 3.12 and 3.13. Python 3.14 has no matching numba wheel, and the pin blocks every newer numba through 0.67.0. - Only two versions are left on PyPI. `pip install pandas-ta==0.3.14b0` — the pin in most code that uses this library — now fails outright, and the two surviving releases need Python 3.12 or newer, so dropping the pin is an interpreter upgrade rather than a one-line edit. - No documentation. `help(ta.macd)` in a REPL is what is left. - No issue tracker, no changelog and no way to report a bug or read why a number looks wrong. - Nothing will be fixed. A dependency break — pandas 3, NumPy 3 — ends the package, and there is no repository from which to patch it. - Indicator accuracy was always "highly correlated with TA-Lib", not identical; with the docs gone, which indicators diverge and by how much is no longer documented anywhere. ## Alternatives **pandas-ta-classic** is the fork the community moved to: MIT, 0.8.32 released 14 September 2026, last commit two days after that, 224 indicators and 62 candlestick patterns implemented natively in Python, and named on ta-lib.org's install page as the pandas wrapper for TA-Lib. It keeps the `.ta` accessor, so it is close to a drop-in. `pandas-ta-openbb` is a narrower fork maintained for [OpenBB](https://stockmarketstack.com/tools/openbb), last released March 2026. For the reference implementation and a maintained C core, [TA-Lib](https://stockmarketstack.com/tools/ta-lib) itself. [VectorBT](https://stockmarketstack.com/tools/vectorbt) computes its own indicators if the surrounding work is a parameter sweep rather than a single series. ## FAQ ### Is pandas-ta still maintained? Not in any way a user can see. The last PyPI release, 0.4.71b0, was uploaded on 14 September 2025, and it and 0.4.67b0 are the only two versions still on the index — every earlier release has been deleted. The GitHub repository has been deleted too, the documentation site no longer answers, and PyPI shows a single maintainer account named pta. The package still installs; nothing else about it is live. ### What happened to the GitHub repository? It was taken down rather than archived. The Internet Archive has a working capture of github.com/twopirllc/pandas-ta on 2 April 2025 and a 404 on 25 June 2025. The account still exists and now lists no public repositories at all, so there is no issue tracker, no history and no fork point. ### What should I use instead? pandas-ta-classic, a community fork by GitHub user xgboosted, is where the activity went — 0.8.32 on 14 September 2026, last commit two days later, MIT, and listed on ta-lib.org's own install page as the pandas wrapper for TA-Lib. It keeps the DataFrame accessor and implements all 62 candlestick patterns in Python. ### Which Python versions does the last release work on? Realistically 3.12 and 3.13. The metadata requires 3.12 or newer and pins numba to exactly 0.61.2, which shipped wheels for CPython 3.10 through 3.13 in April 2025 and none for 3.14. The exact pin also blocks every later numba, up to 0.67.0 from August 2026. ## Also worth comparing - [TA-Lib](https://stockmarketstack.com/tools/ta-lib.md) — The C indicator library everything else wraps — and pip now ships the C part with it. - [arch](https://stockmarketstack.com/tools/arch.md) — GARCH and the rest of the volatility-model family, plus the tests you need around them. - [empyrical-reloaded](https://stockmarketstack.com/tools/empyrical-reloaded.md) — The maintained fork of Quantopian's empyrical — risk statistics, no plots. - [exchange_calendars](https://stockmarketstack.com/tools/exchange-calendars.md) — Sessions, minutes and holidays for 69 exchanges, keyed by ISO-10383 code. - [ffn](https://stockmarketstack.com/tools/ffn.md) — Performance stats, drawdowns and portfolio weights from a DataFrame of prices. - [PyPortfolioOpt](https://stockmarketstack.com/tools/pyportfolioopt.md) — Prices in, weights out — efficient frontier, Black-Litterman and HRP.