Skip to main content

Quickstart

Gomus AI is a cloud-based RAG (Retrieval-Augmented Generation) platform built on deep document understanding. Upload your documents, connect an AI model, and get truthful, citation-backed answers in minutes — no infrastructure required.

This quickstart walks you through:

  • Creating your free account,
  • Choosing an AI model,
  • Uploading documents to a dataset,
  • Starting your first AI chat.

1. Create your account

  1. Go to app.gomus.org and click Sign up.
  2. Register with your email address.
  3. Confirm your email — you are taken to the dashboard.

Every new account starts on the Free plan with 1,000 credits/month. No credit card required.

PlanMonthly creditsKnowledge basesDocs per KBMax file sizeModels
Free1,00025010 MB20 Groq
Base $19.90/mo100,0001050050 MB20 Groq + 22 Bedrock
Premium $49.90/mo250,00020Unlimited200 MB20 Groq + 22 Bedrock
Business $149.90/mo750,000UnlimitedUnlimited500 MB20 Groq + 22 Bedrock
tip

See AI Models in Gomus AI for full details on what each tier includes.

2. Choose your AI model

Gomus AI needs two types of models to work: a chat model (generates answers) and an embedding model (understands your documents).

On the Free plan, you get access to 20 Groq models — fast and high-quality, included in your credit allowance. Paid plans (Base and above) unlock all 22 AWS Bedrock models as well.

To configure your models:

  1. Click your avatar on the top right > Model providers.
  2. The available models for your plan are already listed — no API keys needed.
  3. Click System Model Settings to select your default models:
    • Chat model — we recommend starting with llama-3.3-70b-versatile (1 credit per 1K tokens).
    • Embedding model — select an embedding model for document processing.
note

See Model pricing reference for the complete list of all 42 models with credit costs.

3. Create your first dataset

A dataset (knowledge base) is a collection of documents that Gomus AI parses into searchable chunks. Your AI chat answers will be grounded in these documents.

Supported file formats include PDF, DOC/DOCX, TXT, MD, CSV, XLSX, PPT/PPTX, JPEG, PNG, and more.

To create a dataset:

  1. Click the Dataset tab at the top of the page > Create dataset.

  2. Enter a name for your dataset and click OK.

    You are taken to the Configuration page.

    dataset configuration

  3. Select your embedding model and chunking method (template) for this dataset.

    IMPORTANT

    Once you select an embedding model and parse a file with it, you cannot change it. All files in a dataset must use the same embedding model to ensure consistent results.

  4. Click + Add file > Local files to upload a document.

  5. In the uploaded file entry, click the play button to start parsing:

    parse file

    NOTE
    • If your file parsing gets stuck at below 1%, see this FAQ.
    • If your file parsing gets stuck at near completion, see this FAQ.

4. Review your chunks

Gomus AI gives you full visibility into how your documents are chunked, and lets you intervene where needed.

  1. Click on a parsed file to view its chunks:

    chunks

  2. Hover over each snapshot for a quick preview.

  3. Double-click any chunk to edit it — add keywords or questions to improve retrieval:

    update chunk

    NOTE

    Adding keywords or questions to a chunk increases its keyword weight and improves its position in search results for matching queries.

  4. Use Retrieval testing to verify your setup — type a question in Test text and confirm the results are accurate:

    retrieval test

5. Start your first AI chat

Once your dataset is ready, you can create an AI chat assistant grounded in your documents.

  1. Click the Chat tab at the top of the page > Create chat.

  2. Click the new chat to open its configuration page.

  3. Configure Chat settings on the right side:

    • Name your assistant and select the datasets it should use.
    • Empty response — set a fallback message for when no relevant answer is found in your documents. Leave blank to let the model improvise (may produce hallucinations).

  4. Optionally customise the System prompt.

  5. Select a Chat model from the dropdown (defaults to your system model).

  6. Start chatting:

    chat example

    Gomus AI responds with grounded answers and citations from your documents.

What's next?

You now have a working AI chat assistant backed by your own documents. Here are some next steps to explore: