# Finance ChatGPT Plugins

> Source: https://aiwiki.ai/wiki/finance_chatgpt_plugins
> Updated: 2026-06-23
> Categories: ChatGPT, Finance AI
> From AI Wiki (https://aiwiki.ai), a free encyclopedia of artificial intelligence. Quote with attribution.

*See also: [ChatGPT Plugins](/wiki/chatgpt_plugins), [ChatGPT Plugin Categories](/wiki/chatgpt_plugin_categories) and [Finance](/wiki/finance)*

**Finance ChatGPT Plugins are a now-deprecated category of third-party extensions** that, during 2023 and 2024, connected [ChatGPT](/wiki/chatgpt) to live market data, regulatory feeds, payments rails, and personal finance tools. [OpenAI](/wiki/openai) introduced the plugins protocol on March 23, 2023 as a beta for ChatGPT Plus subscribers, then disabled new plugin conversations on March 19, 2024 and shut down all remaining plugin conversations on April 9, 2024, replacing the program with [Custom GPTs](/wiki/gpts), the [GPT Store](/wiki/gpt_store), and later built-in connectors and [deep research](/wiki/deep_research) [1][2][13]. The plugins beta is therefore historical: there is no live ChatGPT plugin store today, and finance plugins such as Wolfram, Polygon, and FiscalNote no longer run as plugins. Finance was one of the most active plugin categories during the beta, spanning shopping payments, equities and options data, regulatory and political intelligence, and quantitative analysis.

## What were finance ChatGPT plugins?

The finance category sat at the intersection of two design goals OpenAI emphasized when it announced the plugin protocol: extending [large language models](/wiki/large_language_model) past their training cutoff with current, source-of-truth information, and letting users carry out actions that required authenticated APIs. OpenAI described plugins as "tools designed specifically for language models with safety as a core principle" that "help ChatGPT access up-to-date information, run computations, or use third-party services" [1]. For a finance plugin, that meant the model could answer questions like "what is Apple's latest reported revenue" or "what are the top-rated semiconductor stocks today" by fetching data from a real provider rather than guessing from older training data.

A finance plugin typically exposed an `ai-plugin.json` manifest, an OpenAPI specification, and a backend service that served structured responses. ChatGPT used [function calling](/wiki/function_calling)-style routing under the hood to choose when to invoke a plugin, what parameters to pass, and how to summarize the response in the chat. This pattern is the direct ancestor of the **actions** mechanism in Custom GPTs and of broader [tool use](/wiki/tool_use) in modern agentic systems.

## When were ChatGPT plugins launched and deprecated?

Plugins were a roughly one-year beta. The launch shipped with twelve third-party partners plus two OpenAI-hosted plugins (a web browser and a code interpreter) [1]. The program was wound down in March and April 2024 after OpenAI shifted users toward Custom GPTs. OpenAI stated in its help-center notice that "based on the adoption of GPTs by both users and builders, we've decided to wind down the plugin beta" [2][13].

| Date | Event |
|------|-------|
| March 23, 2023 | OpenAI announces the ChatGPT plugins beta with twelve third-party launch partners (Expedia, FiscalNote, Instacart, KAYAK, Klarna, Milo, OpenTable, Shopify, Slack, Speak, Wolfram, Zapier), plus OpenAI's own browsing and code interpreter plugins [1]. |
| May 12, 2023 | Plugins roll out broadly to ChatGPT Plus subscribers. |
| Late 2023 | The plugin store grows past 1,000 plugins; finance and investing become one of the most populated categories. |
| November 6, 2023 | OpenAI introduces Custom GPTs at DevDay, signaling a shift away from plugins [14]. |
| January 10, 2024 | OpenAI launches the GPT Store; users had created more than 3 million custom GPTs in the two months since the DevDay announcement [13]. |
| March 19, 2024 | New plugin installs and new plugin-powered conversations are disabled [2]. |
| April 9, 2024 | All existing plugin conversations are shut down; the plugin platform is fully deprecated [2]. |

## What were the notable finance plugins?

The table below covers plugins that were verifiably published during the plugins beta and that fell within the finance, investing, payments, or regulatory data scopes. Each entry was confirmed against the developer's own announcement, the plugin store listing, or contemporaneous press coverage. All of these plugins ceased operating as ChatGPT plugins when the beta closed on April 9, 2024.

| Plugin | Developer | Function |
|--------|-----------|----------|
| [Klarna](/wiki/klarna) | Klarna | Shopping search and price comparison across hundreds of thousands of retailers, with links into Klarna's pay-later checkout. |
| Wolfram | Wolfram Research | Curated computational knowledge, real-time data, and step-by-step calculations covering currency conversion, interest, present value, and statistics. |
| FiscalNote | FiscalNote | Real-time legal, political, and regulatory data including CQ congressional tracking, Roll Call coverage, and Oxford Analytica briefs. |
| Polygon (later renamed Massive) | Polygon.io | Stocks, options, forex, and crypto market data, ticker search, full option chains, and detailed financial statements. |
| AITickerChat | StockAdvisor.com | Q&A over SEC filings and earnings call transcripts for U.S.-listed equities. |
| Abridged Due Diligence | Independent | Summaries built from recent SEC filings and earnings transcripts to support equity research workflows. |
| Boolio Invest | Boolio | Quantitative factor models, screening, and backtesting on a global equities database. |
| Boo Invest | Boo | Aggregated analyst ratings and price targets from major sell-side firms plus internal quant signals. |

A few other launch-partner plugins, while not strictly finance products, were frequently used in finance-adjacent workflows. KAYAK and Expedia served corporate travel cost estimation, Instacart and Shopify connected to consumer spending and merchant data, OpenTable handled dining reservations and expense planning, and Zapier let analysts pipe ChatGPT outputs into spreadsheets and CRM systems.

### Klarna

[Klarna](/wiki/klarna), the Swedish buy-now-pay-later company, was one of the original twelve launch partners and described itself as among the first brands to build an integrated plugin under the OpenAI protocol [3][6]. The Klarna plugin let users describe a budget or product need in natural language, for example "I have $150, which headphones can I afford," and returned curated product recommendations with deep links into Klarna's compare and checkout tools, which at launch covered roughly 500,000 retail partners [3]. The plugin was rolled out first to U.S. and Canadian ChatGPT Plus subscribers.

### Wolfram

The Wolfram plugin, built on Wolfram|Alpha and the Wolfram Language, gave ChatGPT a numerical and symbolic computation backend. For finance use cases, it handled deterministic math that pure language models often get wrong, including compound interest, bond pricing, internal rate of return, currency conversion at live rates, and statistical analysis of historical price series. It was widely cited in early reviews as one of the most reliable plugins because its responses were grounded in Wolfram's curated data rather than free-form generation. Like every other plugin, it stopped functioning as a ChatGPT plugin after April 9, 2024, though the underlying Wolfram service remains available through other channels.

### FiscalNote

FiscalNote was selected by OpenAI as the inaugural launch partner for legal, political, and regulatory data [4]. The plugin surfaced material from CQ (congressional tracking), Roll Call (legislative news), and Oxford Analytica (geopolitical analysis). Finance users tapped it to monitor proposed legislation that could move particular stocks, sectors, or asset classes, and to brief themselves on policy risk before earnings calls or investment committee meetings.

### Polygon and Massive

The Polygon plugin, from market data provider Polygon.io, exposed the company's full coverage of stocks, options, forex, and crypto through ChatGPT [5]. It supported ticker lookup, real-time and historical pricing, full option chains, news, and the standard income statement, balance sheet, and cash flow statement views. Polygon later rebranded its consumer-facing AI product to Massive, and the same plugin continued to ship under that name through the end of the beta.

### Investing-focused plugins

A second wave of investing plugins appeared after the launch cohort [8].

* **AITickerChat** indexed SEC filings (10-K, 10-Q, 8-K) and earnings call transcripts so users could ask questions like "how did management describe demand softness on the most recent call."
* **Abridged Due Diligence** focused on synthesizing the same primary sources into short briefs for equity research.
* **Boolio Invest** offered quantitative screening and backtesting against a global database of listed equities, including built-in factor models.
* **Boo Invest** aggregated sell-side analyst ratings and price targets and combined them with proprietary quant signals.

These plugins illustrated how third-party finance services used the plugin protocol to convert ChatGPT into a thin natural language interface over their existing APIs.

## What were finance plugins used for?

Finance plugins clustered around four broad jobs.

* **Market data lookup.** Polygon, Wolfram, and the various stock plugins let users ask for live quotes, historical performance, option chains, and financial statements without leaving the chat. This was the single most common workflow, since it directly addressed the [knowledge cutoff](/wiki/large_language_model) problem of static training data.
* **Equity research and due diligence.** AITickerChat, Abridged Due Diligence, and Boolio Invest helped users summarize SEC filings, search transcripts for specific phrases, and screen stocks against quantitative criteria. The pattern foreshadowed retrieval-heavy finance assistants such as AlphaSense and [Hebbia](/wiki/hebbia).
* **Personal finance and shopping.** Klarna covered retail spending decisions, while Wolfram handled budgeting math, loan amortization, and currency conversion. Travel plugins (KAYAK, Expedia) and consumer plugins (Instacart, Shopify) overlapped with personal finance whenever the question was "can I afford this" or "what does this trip cost."
* **Policy and regulatory monitoring.** FiscalNote surfaced legislative and regulatory developments that influence sectors and asset prices, supporting the macro and policy-research side of investment decision making.

## How did the plugin architecture work?

A finance plugin had three required artifacts. First, a plugin manifest at `/.well-known/ai-plugin.json` that described the plugin name, logo, contact information, and authentication scheme. Second, an OpenAPI specification that listed the available endpoints and their parameters. Third, a hosted API that returned JSON responses ChatGPT could read.

Authentication options included no auth (for public data such as exchange rates), service-level API keys, OAuth for user accounts, and user-supplied keys typed into the plugin store. Klarna's product search used service-level auth, FiscalNote used a token-based scheme to gate premium content, and Polygon used user-supplied API keys so each user counted against their own data quota. The pattern of API-key plugins for paid data feeds carried directly into the actions feature in Custom GPTs.

Under the hood, ChatGPT used the same routing primitives as [function calling](/wiki/function_calling): the model decided when a question warranted an external call, formatted the call as a JSON payload that matched the OpenAPI spec, sent the call through OpenAI's plugin runtime, and then summarized the response in the chat. This made plugins an early commercial demonstration of [tool use](/wiki/tool_use) at scale.

## What were the limitations and criticisms?

Finance plugins inherited several constraints that ultimately motivated OpenAI's decision to wind the program down.

* **Discoverability.** With more than 1,000 plugins in the store, users struggled to find the right one for a given question. Finance was crowded, and many plugins offered overlapping capabilities with no clear way to compare them.
* **Single-plugin sessions.** ChatGPT initially limited users to three active plugins per conversation, which made multi-step finance workflows (for example, fetching a price from Polygon, then reasoning about a regulatory filing from FiscalNote, then computing returns in Wolfram) clunky.
* **Hallucinated calls.** The model sometimes invented plausible-looking data instead of calling the plugin, especially for well-known tickers, which created subtle accuracy risks for traders who trusted the chat surface.
* **Compliance ambiguity.** Plugins that surfaced equity ratings, ticker recommendations, or tax-adjacent advice raised questions about whether ChatGPT was acting as an unregistered investment adviser. Most plugins responded by adding disclaimers and avoiding personalized recommendations, but the regulatory perimeter was never formally tested.
* **Limited monetization.** OpenAI did not provide a native billing rail for plugins, so developers either gave functionality away free or maintained their own subscription systems. This dampened investment in long-tail finance plugins.

## What were the regulatory considerations?

Finance plugins operated against a backdrop of unsettled rules for AI in financial services. In the United States, the Securities and Exchange Commission (SEC) proposed conflict-of-interest rules in July 2023 governing how broker-dealers and investment advisers use predictive data analytics, including LLM-based tools [10]. The Financial Industry Regulatory Authority (FINRA) issued a regulatory notice in June 2024 reminding member firms that existing supervision and recordkeeping rules apply when employees use generative AI tools [11]. The European Union's AI Act, adopted in 2024, classifies certain financial AI use cases (such as creditworthiness scoring) as high risk and imposes documentation, monitoring, and human-oversight requirements [12].

Most finance plugins were positioned as informational tools rather than personalized investment advice, and chat outputs typically carried disclaimers reminding users that responses were not recommendations. Even with disclaimers, regulated firms in banking, insurance, and asset management generally used plugins behind their own compliance review processes rather than at the consumer ChatGPT surface, a pattern that has continued with successor systems.

## What replaced ChatGPT finance plugins?

When OpenAI shut down plugins in April 2024, the finance functions migrated along three tracks.

* **Custom GPTs and actions.** Many plugin developers, including FiscalNote and Polygon, rebuilt their integrations as [Custom GPTs](/wiki/gpts) in the [GPT Store](/wiki/gpt_store) using the actions framework, which mapped one-to-one onto the plugin manifest and OpenAPI patterns. Actions also supported richer authentication flows and could be embedded into multi-tool GPTs.
* **API-level integrations.** Some teams shifted from end-user plugins to direct integrations through the OpenAI API and [function calling](/wiki/function_calling), giving them tighter control over prompts, retrieval, and compliance logging. This is the dominant pattern for regulated finance firms today. OpenAI later folded similar capabilities back into ChatGPT itself through built-in browsing, connectors, and [deep research](/wiki/deep_research), reducing the need for separate finance plugins.
* **Vertical AI products.** Independent finance-AI vendors expanded rather than relying on ChatGPT's plugin surface. Bloomberg announced BloombergGPT, a 50-billion parameter finance-domain model, in March 2023 [9]. AlphaSense built retrieval and summarization for analyst workflows on top of its own search index, and [Hebbia](/wiki/hebbia) developed agentic research over private documents. These products generally outperformed plugin-based workflows on professional finance tasks because they could be tuned for accuracy, citation, and compliance.

## See also

* [ChatGPT Plugins](/wiki/chatgpt_plugins)
* [ChatGPT Plugin Categories](/wiki/chatgpt_plugin_categories)
* [Finance](/wiki/finance)
* [AI in finance](/wiki/ai_in_finance)
* [Custom GPTs](/wiki/gpts)
* [GPT Store](/wiki/gpt_store)
* [Function calling](/wiki/function_calling)
* [Tool use (artificial intelligence)](/wiki/tool_use)
* [Deep research](/wiki/deep_research)
* [OpenAI](/wiki/openai)
* [ChatGPT](/wiki/chatgpt)
* [Hebbia](/wiki/hebbia)
* [Cryptocurrency ChatGPT Plugins](/wiki/cryptocurrency_chatgpt_plugins)

## References

1. OpenAI. "ChatGPT plugins." March 23, 2023. https://openai.com/index/chatgpt-plugins/
2. OpenAI Developer Community. "Plugin Store and New Chats With Plugins - Closed March 19 2024." https://community.openai.com/t/plugin-store-and-new-chats-with-plugins-closed-march-19-2024/689877
3. Klarna. "Klarna brings smoooth shopping to ChatGPT." PR Newswire, March 23, 2023. https://www.prnewswire.com/news-releases/klarna-brings-smoooth-shopping-to-chatgpt-301780170.html
4. FiscalNote. "FiscalNote Selected by OpenAI for Collaboration As Inaugural Launch Partner for OpenAI's ChatGPT Plug-in." Business Wire, March 23, 2023. https://www.businesswire.com/news/home/20230323005565/en/
5. Polygon.io / Massive. "Announcing the Polygon ChatGPT Plugin." https://massive.com/blog/polygon-plugin-for-chatgpt
6. TechCrunch. "Klarna plugs ChatGPT into its platform for faster product recommendations." March 23, 2023. https://techcrunch.com/2023/03/23/klarna-chatgpt-shopping/
7. Zapier. "Sunsetting the Zapier ChatGPT plugin: what you need to know." https://help.zapier.com/hc/en-us/articles/24785309335565
8. Prompt Engineering Institute. "ChatGPT Plugins: the Future of Personal Finance and Investing?" https://promptengineering.org/the-proliferation-of-finance-and-investing-chatgpt-plugins/
9. Bloomberg. "Introducing BloombergGPT, Bloomberg's 50-billion parameter large language model, purpose-built from scratch for finance." March 30, 2023.
10. U.S. Securities and Exchange Commission. "Proposed Rule: Conflicts of Interest Associated with the Use of Predictive Data Analytics by Broker-Dealers and Investment Advisers." Release Nos. 34-97990; IA-6353. July 26, 2023.
11. FINRA. "Regulatory Notice 24-09: Firms' Use of Generative Artificial Intelligence and Large Language Models." June 27, 2024.
12. European Parliament. "Regulation (EU) 2024/1689 (AI Act)." 2024.
13. OpenAI. "Introducing the GPT Store." January 10, 2024. https://openai.com/index/introducing-the-gpt-store/
14. OpenAI. "New models and developer products announced at DevDay." November 6, 2023. https://openai.com/index/new-models-and-developer-products-announced-at-devday/