A branching research workspace

It’s 9pm. Alex has a literature review due Monday and forty tabs open.

This is what happens when she asks fork.ai instead.

No signup needed for a first session
Alex’s first move

A chat tab, then a mess

Alex’s first move was the obvious one — a chat tab. The first answer was great. By message thirty she couldn’t tell which reply was about socioeconomic moderators and which was about measurement methodology. Scrolling up cost her more time than writing would have.

Linear chat
green space + mental health, moderating factors?
Socioeconomic status significantly moderates…
wait, what about measurement methods?
Good question — most studies use self-report…
going back to the moderators thing —
⋯ 30 messages later, still scrolling
fork.ai
Root question
Moderating factors
Measurement methods
Open questions

Every thread Alex opens becomes its own node — so she can go as deep as she wants on one idea without losing the rest.

Alex asks a real question

One question, split into sections

Not one long blob of text — an answer she can navigate, one thread at a time.

“How does access to urban green space affect mental health outcomes?”
1 · Overview

Across dozens of observational studies, access to urban green space is consistently associated with lower rates of depression, anxiety, and perceived stress — though effect sizes vary widely by city and how “access” itself is measured.

2 · Moderating factors

Socioeconomic status significantly moderates this relationship — lower-income residents see larger mental health gains from nearby green space, plausibly because they have fewer alternative ways to decompress.

Try selecting a sentence above ↑
3 · Measurement challenges

Studies disagree on whether “access” means walking distance, canopy cover from satellite imagery, or self-reported visit frequency — and the choice changes the measured effect size substantially.

4 · Open questions

Causality is still contested: does green space improve mental health, or do mentally healthier people simply seek it out more? Few studies track the same residents over time to settle this.

One click, and “moderating factors” becomes its own node — with its own clean context, not a jumble of everything Alex’s asked so far.

Zoom out

Every branch becomes a node

Root questionModerating factorsAsk AI · highlighted sentence+ more as she explores

Two clean branches already, each traceable back to the exact question or sentence that spawned it. Nothing buried, nothing blended together.

Alex’s next move

Pick a model per branch

The moderating-factors branch matters for her defense, so she wants something sharper. A quick side-question stays on the cheap default — she’s paying per branch, not per session.

Claude Haiku
~1× the cost of the cheapest tier

The first question always uses a fast default model — you choose the model for what you branch into.

One more branch

Optional web search, with real sources

Recent papers, not just what the model already knew — flip it on per branch.

Recent meta-analyses continue to support a moderate association between green space access and reduced depressive symptoms, with socioeconomic status as a consistent moderator.

Sunday night wraps

Push the whole map into Notion

One click, and the whole session lands as a real Notion page — not a tab she’ll never reopen.

Root questionModerating factorsAsk AI
Mind map (auto-generated)

Access to green space is broadly associated with better mental health outcomes across dozens of observational studies.

It slots into the thesis doc Alex already has open, instead of living in a tab she’ll never reopen.

Monday morning

Share the map, no account needed

Alex sends the link to her advisor, who opens it and branches straight away — no login, no waiting.

AL
AD
Alex · Advisor (guest, no account)

If the advisor signs up later, anything they added on the shared map is automatically theirs. First session is free — up to 5 nodes before anyone’s asked to log in.

Who it’s for

Alex isn’t the only one

Students

Coursework, problem sets, and papers due at 11:59pm. Ask, branch, and keep what matters — without forty browser tabs.

Researchers

Literature reviews, methodology deep-dives, comparing papers. Keep every thread traceable back to the exact question or quote that started it.

Notion & PKM users

Already living in Notion, Obsidian, or a Zettelkasten? Research sessions land as structured pages and diagrams — not a transcript you’ll never reopen.

Alex’s other question

Don’t pay a flat fee. Pay for what you use.

She used about 15 branches that week, mixed cheap and pricier models. Would a flat monthly plan have been worth it — or is pay-as-you-go cheaper for usage that’s bursty around deadlines? Try your own numbers below.

Typical model tier
Claude Haiku tier — Gemini, DeepSeek, and GLM are typically cheaper still.
$/month
$20 is just a common flat-fee reference point for illustration — edit it to match whatever you’re comparing against.
About $0.93/month with fork.ai’s pay-as-you-go credit, at 40 questions/month on the fast tier.
That’s about $19.07 less than a flat $20.00/month plan.
Your free $5.00 signup credit covers about 215 questions at this tier before you’d need to add credit.

Estimate based on a typical multi-section answer (~1,500 input / ~2,800 output tokens) — your mileage will vary with question complexity and model choice. Not a guarantee.

Questions

Frequently asked

No — your first session is free, up to 5 nodes, no account needed. Sign up later and any guest branches you made are automatically yours.

Ask once. Branch forever.

Try fork.ai free