# AI use cases for Rizzo Flow (Rizzo-AI-Academy/rizzo-flow) Repo: [Rizzo-AI-Academy/rizzo-flow](https://github.com/Rizzo-AI-Academy/rizzo-flow) | Use cases: 4 | Page: https://dedupedai.com/use-cases/rizzo-ai-academy-rizzo-flow/ Rizzo Flow is an open, local-first tool that returns typed, probabilistic decisions (yes/no, choices, rubric scores, numbers) from unstructured state using a local LLM over an HTTP API. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Get typed LLM decisions from text on your own machine: https://dedupedai.com/use-cases/rizzo-ai-academy-rizzo-flow/local-typed-decisions-api/ - Triage a support ticket from the command line: https://dedupedai.com/use-cases/rizzo-ai-academy-rizzo-flow/triage-ticket-from-command-line/ - Run existing TypeSafe Jev API code against localhost: https://dedupedai.com/use-cases/rizzo-ai-academy-rizzo-flow/point-jev-code-at-localhost/ - Run a smaller or quantized model on modest hardware: https://dedupedai.com/use-cases/rizzo-ai-academy-rizzo-flow/run-smaller-model-on-weak-hardware/ ## How to get typed LLM decisions from text on your own machine with Rizzo Flow Repo: [Rizzo-AI-Academy/rizzo-flow](https://github.com/Rizzo-AI-Academy/rizzo-flow) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Run a local decision API and get yes/no, choice and score answers as JSON with probabilities. **Who it is for:** A developer who wants an LLM to classify or score messages, tickets or records, and does not want the data to leave their machine. **The problem:** LLM answers are free text that must be parsed, and hosted decision APIs send your data to a third party. **The outcome:** A running local server at http://127.0.0.1:8017 that returns typed decisions with probabilities for any state you send. ### Steps 1. Install Python 3.11 or newer, git and uv. 2. Run `git clone https://github.com/Rizzo-AI-Academy/rizzo-flow && cd rizzo-flow` 3. Run `uv sync --locked` 4. Run `uv run rizzo download` and wait for the model and runtime to finish. 5. Run `uv run rizzo serve`, then open http://127.0.0.1:8017/playground to try a state with questions. 6. Send the curl example from the Quickstart to http://127.0.0.1:8017/v1/systemone with your own state and questions. ### Prompt to paste into your agent ```text In the rizzo-flow repo, start the server with `uv run rizzo serve`, then POST a JSON body with the state from and one yes/no question to http://127.0.0.1:8017/v1/systemone using model rizzo-latest. Return the typed decision and its probabilities. Verify that the response contains a probability for each option. ``` ### Caveat Probabilities are uncalibrated unless you calibrate them on your own data, and the fine-tuned model is still below TypeSafe Jev in accuracy (0.648 against 0.727 on the typed-decisions benchmark in the README). 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/rizzo-ai-academy-rizzo-flow/local-typed-decisions-api/ ## How to triage a support ticket from the command line with Rizzo Flow Repo: [Rizzo-AI-Academy/rizzo-flow](https://github.com/Rizzo-AI-Academy/rizzo-flow) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Run one decision script on a ticket file without starting a server. **Who it is for:** A support or operations person who has a ticket saved as a JSON file and wants a quick typed answer such as whether it is urgent. **The problem:** Reading every ticket by hand to set priority takes time, and starting a server for one check is extra work. **The outcome:** A typed decision printed for one ticket file, such as an urgency answer with its probability. ### Steps 1. Install Python 3.11 or newer, git and uv. 2. Run `git clone https://github.com/Rizzo-AI-Academy/rizzo-flow && cd rizzo-flow` 3. Run `uv sync --locked` and then `uv run rizzo download`. 4. Write your ticket and its questions into a JSON file in the same format as examples/ticket.json. 5. Run `uv run rizzo decide examples/ticket.json`, replacing the path with your own file. ### Prompt to paste into your agent ```text In the rizzo-flow repo, write a JSON file at with the ticket text as state and an is_urgent question, then run `uv run rizzo decide `. Return the decision and its probability. Verify that the output is a typed answer and not free text. ``` ### Caveat The README says probabilities are uncalibrated, so treat the urgency score as a ranking aid and check it against your own tickets before relying on it. 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/rizzo-ai-academy-rizzo-flow/triage-ticket-from-command-line/ ## How to run existing TypeSafe Jev API code against localhost with Rizzo Flow Repo: [Rizzo-AI-Academy/rizzo-flow](https://github.com/Rizzo-AI-Academy/rizzo-flow) | Difficulty: Advanced | Time: 60 minutes | Basis: From the README Change one URL so code written for the TypeSafe API calls a local model instead. **Who it is for:** A developer whose app already calls the TypeSafe Jev API and who wants to test it offline or avoid hosted calls. **The problem:** Hosted Jev calls cost money per use and send data to a closed service, which blocks some local or private testing. **The outcome:** Your existing Jev-style code returns decisions from a model on localhost after a one-line base URL change. ### Steps 1. Set up the server by running `uv sync --locked`, `uv run rizzo download` and `uv run rizzo serve` from the Rizzo Flow folder. 2. If you need Bearer auth on the Jev-compatible endpoints, set RIZZO_API_KEY before starting the server. 3. In your app, change the TypeSafe API base URL to http://127.0.0.1:8017. 4. Run your existing tests or a sample call and compare the returned decisions with your expected results. 5. Note that the base model answers 'cannot determine' often with abstention on, so the Jev-compatible endpoint always disables abstention. ### Prompt to paste into your agent ```text In the rizzo-flow repo, start the server with `uv run rizzo serve`, then point the existing TypeSafe API client in at http://127.0.0.1:8017 and run its test suite. Return the failing or passing tests. Verify that each decision response has the same shape the client expects. ``` ### Caveat Rizzo Flow is not affiliated with TypeSafe and does not reproduce Jev's architecture, so expect different answers and probabilities from the hosted service. 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/rizzo-ai-academy-rizzo-flow/point-jev-code-at-localhost/ ## How to run a smaller or quantized model on modest hardware with Rizzo Flow Repo: [Rizzo-AI-Academy/rizzo-flow](https://github.com/Rizzo-AI-Academy/rizzo-flow) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Pick a lighter model size or quantization to run decisions on a laptop with limited memory. **Who it is for:** A person with a laptop or a CPU-only machine who wants to try typed decisions without the 4B default model. **The problem:** The default 4B Q8_0 model is about 4.4 GB, which may be too large or too slow on a weak machine. **The outcome:** A running server with a smaller model, and a measured comparison of its answers against the default on your own workload. ### Steps 1. Run `uv sync --locked` from the Rizzo Flow folder. 2. Run `uv run rizzo devices` to see what the runtime detects and which device `--device auto` will use. (inferred, check the README) 3. Run `uv run rizzo download --size 1.7b` for a faster, smaller first look. 4. Run `uv run rizzo serve --size 1.7b` or `uv run rizzo serve --quant q4_k_m` to choose the model and quantization. (inferred, check the README) 5. Download the Q4_K_M files first with `uv run rizzo download` if the quantization is not present yet, then run the same test questions on each option. (inferred, check the README) ### Prompt to paste into your agent ```text In the rizzo-flow repo, run `uv run rizzo devices`, then start the server with `uv run rizzo serve --size 1.7b` and send the same three questions from that were answered by the default model. Return the answers and probabilities side by side. Verify that accuracy differences are recorded before choosing a model. ``` ### Caveat The README says the 1.7B model is much less accurate and that quantization and hardware change probabilities, so compare results on your own data before you choose one. 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/rizzo-ai-academy-rizzo-flow/run-smaller-model-on-weak-hardware/