A good summary preserves the argument while cutting the volume. This tool takes any document — an article, a report, a contract, a research paper — and uses your own OpenAI or Anthropic key to compress it into a faithful summary at the length and format you choose. Everything runs in your browser; your key never touches a Gera server.
How it works
Choose a provider and model, paste your API key, drop in your document, and pick a summary length (short, medium, or detailed) and a format (paragraph, bullets, or TL;DR). The tool builds a structured summarisation prompt that tells the model exactly how long the output should be, how to shape it, and — critically — to stay faithful to the source without inventing facts. It then sends one direct request to the provider and returns the summary for you to copy.
For Anthropic, the request includes the official direct-browser-access header so the call works straight from the page.
When to use which length
- Short — a 2-3 sentence gist for triage or a Slack message.
- Medium — a one-paragraph abstract for a reading list or email digest.
- Detailed — 2-3 paragraphs that retain nuance for review or hand-off.
Choosing the right output format
The three formats serve different downstream uses:
Paragraph gives you flowing prose that reads naturally. It is the right choice when you need to share the summary in a context where lists look out of place — an email update, an executive brief, or a report appendix. Paragraph summaries also tend to preserve more of the document’s logical flow.
Bullets give you a scannable list of key points. Each bullet is a distinct claim or finding, making it easy to skim for the one point you need. Use this format when you want to review a document quickly for relevance, or when your downstream use is a slide or a brief where hierarchy matters more than narrative.
TL;DR leads with a single-sentence or two-sentence takeaway and follows with supporting bullets. This is the most efficient format for a first-pass triage read, or for sharing in a channel where people may only read the first line.
Handling long documents
The model’s context window is the practical ceiling on a single summarisation call. For most shorter documents — articles, reports up to ten or twenty pages, contracts — a single call is fine. For longer documents, the practical approach is to split the document into logical sections (chapters, major sections, or fixed-length chunks), summarise each section separately, and then run a second summarisation call on the combined section summaries. This “summarise the summaries” approach scales to very long documents and keeps each individual call within context limits.
For contracts and legal documents, it is often more useful to extract specific fields (parties, dates, obligations) rather than summarise the whole document — the Contract Summary Prompt Builder is designed for that task.
Privacy and cost
Your API key is sent only to the model provider — OpenAI or Anthropic — and only for the specific request you trigger. It is not stored, logged, or transmitted to any Gera server; it lives in your browser tab and is cleared on refresh. The text you paste is sent as part of the API request to the provider you select; check that provider’s data handling terms if you are processing sensitive content.
The cost of a typical summarisation call is very small at the per-token rates of current models. Cheaper, faster models — the mini and haiku tiers — are more than capable for summarisation tasks and cost a fraction of flagship model calls.
Tips
- For documents longer than the model’s context window, summarize sections separately, then summarize the combined summaries.
- The cheaper models are more than capable of summarisation — there is rarely a reason to pay for a flagship model here.
- Always sanity-check numbers, names, and dates against the original; summarisation is faithful but not infallible.
- The tool is free — your only cost is the tokens consumed on your own API key.