Every AI interaction in Eden uses "requests" from your monthly allowance. This guide explains how requests work, what affects usage, and how to get the most out of your plan.

The Simple Version

2 requests ≈ 1 message with a long article's worth of context

Most simple messages use about 2 requests. Longer conversations with more context use more. That's really all you need to know — just use Eden and the request counter will show you what you're using.

If you want the details, keep reading.

What Counts as a Request

Requests are used when:

• You send a message in an AI chat — The AI processes your message plus any connected content

• You transcribe a video or podcast — Longer content uses more requests

What Affects Request Usage

Three main factors determine how many requests a message uses:

1. Context Size

The more content the AI needs to process, the more requests it uses. A quick question with no files attached costs less than analyzing a 50-page PDF.

2. Conversation Length

Longer conversations use more requests per message because the AI remembers your full conversation history.

3. Model Tier

More powerful AI models use more requests:

See Working with AI for guidance on choosing models.

What Are Tokens?

Behind the scenes, AI measures everything in "tokens" — small chunks of words. You don't need to count tokens (Eden handles that), but here's the scale:

On average: 1 token ≈ ¾ of a word, so 100 words ≈ 130 tokens.

The more context you give the AI (documents, conversation history, etc.), the more tokens are involved, and the more requests it uses.

How Eden Is Different

Eden isn't a typical AI chatbot, and that changes how you'll use requests.

With most AI tools, you're constantly typing questions and copying/pasting context into every message. It's a back-and-forth where you do most of the work explaining things repeatedly.

With Eden, you work differently:

You're transcribing and analyzing media

Most Eden users spend requests on video and podcast transcription — turning hours of content into searchable, actionable text. You're not just chatting; you're processing real work.

You're referencing your own content

Eden connects to your files, documents, and links. Instead of copying and pasting context into every message, Eden already knows your stuff. That means richer responses without the hassle.

Your chats are deeper, not longer

Because Eden already has context, you don't need to send 50 short messages explaining background. You send fewer, more meaningful messages — and get better answers.

What this means: You might use fewer total messages than with a typical chatbot, but each message does more work.

Requests by Plan

Video & Podcast Capacity by Plan

Real-World Examples

Researcher reading academic papers

• Upload and discuss 5 papers/week → ~60 requests

• Deep analysis and note-taking → ~100 requests

• Follow-up questions and synthesis → ~90 requests

• Monthly total: ~250 requests → Starter covers it

Podcaster with a weekly 1-hour show

• Transcribe 4 episodes/month → ~200 requests

• Generate show notes and timestamps → ~50 requests

• Create clips and social posts → ~100 requests

• Write companion newsletter → ~100 requests

• Monthly total: ~450 requests → Starter covers it

Content creator / Writer

• Daily brainstorming and drafts → ~300 requests

• Research with source documents → ~400 requests

• Editing and feedback → ~200 requests

• Repurposing across platforms → ~200 requests

• Monthly total: ~1,100 requests → Creator covers it

YouTuber transcribing back catalog

• Transcribe 20 videos (30 min each) → ~520 requests

• Generate descriptions and chapters → ~200 requests

• Create blog posts from transcripts → ~400 requests

• Social media clips and posts → ~300 requests

• Monthly total: ~1,420 requests → Creator covers it

Agency managing multiple clients

• Heavy daily use across 5+ projects → ~3,000 requests

• Video/podcast processing → ~1,500 requests

• Document analysis and research → ~2,000 requests

• Monthly total: ~6,500 requests → Pro covers it

Checking Your Usage

To see how many requests you have remaining:

1. Go to Settings

2. Select Billing

3. View your Workspace usage

You'll see requests used, requests remaining, and when your allowance resets.

What Happens When You Run Out

If you hit your request limit, you have two options:

1. Upgrade Your Plan

Move to a higher tier for more monthly requests. See Plans & Pricing for details.

2. Add More Requests

Purchase an add-on pack of 1,500 requests anytime:

Each add-on gets you:

• 750 quick messages, OR

• 375 document-assisted conversations, OR

• 30 one-hour podcast transcriptions

No commitment — just add when you need it.

Tips to Maximize Your Requests

Use Base models for most work

Base models handle 80%+ of tasks well and use the fewest requests. Save Pro and Expert for complex analysis.

Be specific in your questions

A focused question gets a better answer in fewer back-and-forths.

Connect only what you need

On Canvas, connect the specific items relevant to your question rather than everything you have.

Let conversations end naturally

Start a new chat for new topics rather than continuing a very long conversation (which accumulates context).

Common Questions

Why did one message use more requests than another?

The AI processes all the context you've connected. More content (longer videos, bigger documents, longer conversation history) = more requests per message.

Do connected items use requests when I connect them?

No. Connecting an item doesn't use requests. Requests are only used when you send a message and the AI processes your content.

Does transcription count toward my request limit?

Yes. Transcribing videos and podcasts uses requests based on the length of the content.

What if I run out mid-month?

You can purchase an add-on pack instantly or upgrade your plan. There's no waiting.

Do unused requests roll over?

No, requests reset each billing cycle.