A project to build with System One models

Intent detection for a chatbot in your own language

Detect what users want in Thai, Portuguese, Hindi or Nepali without translating first, and route them to the right flow.

MultilingualIntermediate4 stepsSuggested by @biplov

The decision

A choice over your bot's intents, such as check order, refund, change address, talk to a person. The options can be written in English while the user writes in their own language.

How to build it

List your intents with a one-line description each.
Pick a multilingual model. OpenThai-SystemOne is trained for Thai and English; Julia 1 and Laya's multilingual checkpoint cover many languages.
Run 50 real user messages per language through it, and check the accuracy per language before launch.
Below a threshold, ask a clarifying question instead of guessing.

Make it better

Measure each language separately. A model that is strong in English can be much weaker in another language, and a small fine-tune per language fixes most of it.

More projects to build

All projects

MultilingualIntermediate

Support triage for customers who write in Nepali

Route support messages written in Devanagari, Romanized Nepali or a mix with English to the right team, with a model that reads Nepali as people write it, and hand the unclear ones to a person.

sagea/mira convai-innovations/laya supersonic-labs/julia-1+1

LLM routing and guardrailsIntermediate

A guardrail that checks every tool call before an agent runs it

Ask a decision model "is this command safe to run?" before a coding or ops agent executes it, and block or ask for approval below a threshold.

jared-palmer/kev mapika/decider alibi-serikbay/jevk5

LLM routing and guardrailsIntermediate

Route each prompt to the cheapest LLM that can answer it

A decision model reads the prompt and picks the smallest model likely to handle it, cutting LLM spend without hurting quality.

vllm-semantic-router/decision jared-palmer/kev convai-innovations/laya