# AI use cases for Qanat (fidetolabs/qanat) Repo: [fidetolabs/qanat](https://github.com/fidetolabs/qanat) | Use cases: 3 | Page: https://dedupedai.com/use-cases/fidetolabs-qanat/ Qanat is a Python data platform for investment research that screens companies on the numbers they had published as of any past date and serves that data to AI agents over MCP. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Screen companies as they stood on a past date: https://dedupedai.com/use-cases/fidetolabs-qanat/screen-companies-as-of-past-date/ - Measure what screened companies did between filings: https://dedupedai.com/use-cases/fidetolabs-qanat/measure-returns-between-filing-dates/ - Ask an AI agent questions over dated filings: https://dedupedai.com/use-cases/fidetolabs-qanat/ask-agent-questions-over-dated-filings/ ## How to screen companies as they stood on a past date with Qanat Repo: [fidetolabs/qanat](https://github.com/fidetolabs/qanat) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Run a metric threshold screen on SEC filings using only what was public on a chosen date. **Who it is for:** An investment researcher who wants to test a valuation rule without look-ahead bias, after reading that most screeners only show today's data. **The problem:** Standard screeners reflect current data, so a rule tested on past dates quietly uses figures that were published later. **The outcome:** A list of companies that passed the rule on the chosen date, plus a count of how many could and could not be evaluated. ### Steps 1. Install Qanat with `uv tool install qanat-fdtl` (or `pip install qanat-fdtl`). 2. Set your contact address with `export SEC_USER_AGENT="your name you@example.com"`. 3. Create a demo project with `qanat init my-research --demo && cd my-research`. 4. Validate the project file with `qanat check`. 5. Run the screen with `qanat screen --filter '[{"metric":"pe","op":"<","value":30}]' --as-of 2025-12-31`. ### Prompt to paste into your agent ```text In the Qanat project at , run a screen for metric pe below 30 as of using the qanat screen command. Return the companies that passed, the count evaluated, and the count that could not be evaluated. Verify that the list is shown together with the unevaluated count so no companies are silently dropped. ``` ### Caveat Qanat screens only on what companies reported in filings, and the bundled demo has no prices, so it cannot yet show what a list did next. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/fidetolabs-qanat/screen-companies-as-of-past-date/ ## How to measure what screened companies did between filings with Qanat Repo: [fidetolabs/qanat](https://github.com/fidetolabs/qanat) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Run a screen over a date window and read each passing company's return between its own filing dates. **Who it is for:** A researcher who has a screen rule and wants to see how the companies that passed performed over the following filing periods. **The problem:** Companies file on different calendars, so comparing their results over one shared period is unreliable and easy to get wrong by hand. **The outcome:** For each date in the window, the companies that passed, each with its return from its own filing to its next filing, and the evaluated and unevaluated counts. ### Steps 1. Install Qanat with `uv tool install qanat-fdtl` (or `pip install qanat-fdtl`). 2. Set your contact address with `export SEC_USER_AGENT="your name you@example.com"`. 3. Create a demo project with `qanat init my-research --demo && cd my-research`. 4. Run the windowed screen with `qanat screen --filter '[{"metric":"pe","op":"<","value":30}]' --from 2021-12-31 --to 2025-12-31`. 5. Read the per-company returns and the counts, then decide any averaging or position sizing yourself outside Qanat. ### Prompt to paste into your agent ```text Using the Qanat project at , run a windowed screen from to on the metric and threshold . Report the per-company return from each company's own filing to its next filing, along with the evaluated and unevaluated counts. Verify that no average or aggregate has been added, since Qanat reports per-company returns only. ``` ### Caveat Qanat does not add up returns or compare against failed companies, so any portfolio-level result must be built outside the tool. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/fidetolabs-qanat/measure-returns-between-filing-dates/ ## How to ask an AI agent questions over dated filings with Qanat Repo: [fidetolabs/qanat](https://github.com/fidetolabs/qanat) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Connect an MCP client to Qanat's read-only research scope and query SEC filings and screens in plain language. **Who it is for:** An analyst who wants an AI agent to read company data and screen results directly, without writing queries. **The problem:** Pulling figures from many 10-K filings by hand is slow, and an agent without dated data may answer using information that was not yet public. **The outcome:** An agent that answers questions about what data is in the project, where it came from, and which companies passed a rule in a given year, using only dated data. ### Steps 1. Install Qanat with `uv tool install qanat-fdtl` (or `pip install qanat-fdtl`). 2. Set your contact address with `export SEC_USER_AGENT="your name you@example.com"`. 3. Create a demo project with `qanat init my-research --demo && cd my-research`. 4. Register the server with your agent using `claude mcp add qanat -- qanat mcp --scope research`, or add the Qanat entry with cwd set to the project folder in any MCP client. 5. Ask the agent: "Which companies passed my rule in 2022?" ### Prompt to paste into your agent ```text Using the qanat MCP server in with research scope, answer this question: . Reply with the answer, the source filings or tables used, and the dates they were public. Verify that every number comes from a Qanat tool call and not from memory. ``` ### Caveat The dates protect a step but not the person or agent asking, so the agent can still see later data while exploring, and results must be checked for timing. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/fidetolabs-qanat/ask-agent-questions-over-dated-filings/