This guide gets you from zero → insights in minutes.
1) Login and open quick start
- Open the Dashboard
Log in to your account and open the dashboard.
- Click on Quick Start
Look in the sidebar and click on Quick Start.
- Set up your Project specifications
Start by giving your project a name and select a relevant workspace. By default, there are no workspaces. Click the + button to create a new workspace. Make sure to select the correct context. description and excluded lexicon are optional, but let you refine the results. Once ready, click Next.
Context shapes which metrics and suggestions you’ll see. For example, “Marketing — Web” enables SEO & CTA guidance, while Fiction prioritizes style/structure over marketing tips.
2) Upload your first dataset
- ✓
Choose whether you want to bring your own data or have a corpus automatically created. For this example, we will bring our own data.
- ✓
Choose text input or local file upload.
- ✓
Add your data or upload a file in the input section. Click Next.
- ✓
Review your project specifications. If right, click Create & Generate.
We preprocess automatically: language detection, normalization, tokenization, lemmatization, and stopword removal.
3) Run your first analysis
Your analysis will automatically generate and you'll be taken to the overview dashboard. Along the top you’ll see multiple tabs. Here’s how to read them:
- Overview — overview of the dataset, including frequency charts for dominant terms per corpus (with z-scores/keyness).
- Concordance (KWIC) — the contexts a term appears in.
- Collocates — search for word collocations and sort by PMI, t-score, logDice, and LLR.
- Sentiment — positive / negative skew and the token drivers.
- Messaging — clustered themes and top terms per cluster.
- Readout - a one pager for download as a PDF with high value callouts and recommendations (for marketing).
- Options - Corpus options, including any changes you want to make, raw data downloads, and a recompute button.
High resonance means two corpora share language strongly; low resonance highlights differentiation. See Resonance Explained for the interpretation guide.
Frequently asked
Do I need technical skills?
No — everything runs in the browser.
How big can datasets be?
As long as it fits within your plan's token limit, you can make the dataset as large as you want it to be.
Can I analyze non-English?
Yes. We handle normalization, tokenization, and stopwords for multiple languages. For a full list of languages, see Supported Languages.
If your CSV doesn’t parse: ensure there’s a text column, UTF-8 encoding, and no giant cells (>100k chars). If needed, split files and upload in batches.
Next steps
- Compare two real-world corpora (e.g., brand A vs. brand B press, or customer reviews by segment).
- Read the deep-dive: Resonance Explained.
- Explore the Messaging tab to map voice & gaps. (Guide coming soon.)