§ s support

Help, written
by people who run regressions.

QuantFit support is staffed by applied econometricians, not a triage team. If your Hessian is singular at 11pm on a Sunday, an actual person who knows what that means will get back to you.

Channels & response times

How fast you hear back, by tier.

No tier waits longer than three working days for a real human reply. Site-licensed labs have a named CSM and a same-day SLA.

Table s.1 · Support tiers

weekday hours · UK / EU base · global Lab cover

Tier	Channel	Response time	Best for
Student · Free	Community forum	Best-effort · usually 1–3 days	Open-source projects, classroom use, replication assignments.
Researcher · paid	Email — [email protected]	One business day · UK / EU hours	Active subscribers. Methodological questions, bug reports, dataset issues.
Lab · institutional	Named CSM + email	Same business day · global hours	Site-licensed departments. Procurement, custom rollout, training.

SLA windows are measured during UK working hours, Mon–Fri.* Lab tier covers global business hours via the named CSM.

Table s.1 — Channels by tier. Site-licensed labs may also escalate via the shared engineering channel; ask your CSM.

§ c contact

Three ways
to reach us.

For most things, email is fastest. For methodology you’d like to think about in public, the forum has the best signal — most active econometricians on staff read it daily.

✓One-business-day reply on Researcher
✓Same-day reply on Lab
✓Bug reports get a triage number you can quote
✓Methodology questions get attached to the relevant chapter in the manual

General support

[email protected]

Bug reports, methodological questions, dataset issues, billing. Include your workspace ID (in the URL) and the reproducibility hash from the run if it’s about a specific fit.

For labs & sales

[email protected]

Site licenses, procurement, named-user onboarding, data-residency arrangements. See For labs for what we offer institutions.

Forum & community

Open the manual’s discussion thread

Each chapter of the documentation has an open discussion at the bottom. Most useful methodology threads end up cited in future versions of the manual.

App Store review essentials

Support, legal, purchases, and user controls.

These are the user-facing routes Apple reviewers expect to find from support, settings, metadata, or the subscription flow.

What data QuantFit collects, how uploaded datasets are handled, and how users contact us.

Rules for accounts, subscriptions, estimator outputs, downloads, and research use.

Delete Account

In-app deletion path plus email fallback for users who cannot access the app.

Restore Purchases

Instructions for App Store subscription and purchase restoration.

Data & Storage

Plain-language handling of datasets, workspaces, generated outputs, and deletion.

Estimator SOP PDFs

Downloadable SOPs for every estimator, including the combined PDF set.

§ faq

Questions our
methodologists answer.

Pricing FAQ lives on the pricing page; this one is for the model itself.

My fit returned a singular Hessian — what now?: That usually means a perfectly collinear regressor or a degenerate panel structure. QuantFit will name the offending column in the diagnostic panel; drop it or use a within-transform. For PMG-ARDL specifically, check that T exceeds the ARDL lag order in every panel.
My panel diagnostic says "cross-sectional dependence" — does that invalidate my model?: Not necessarily — it means the i.i.d. SE assumption is wrong. Re-run with cluster-robust SE or switch to CS-ARDL (in 2026.1). The Pesaran CD test is reported alongside every panel regression so reviewers see your diagnostic up front.
Can I export the exact spec a colleague used so I can reproduce her result?: Yes. Every fit writes a reproducibility hash and a `.qfspec` file. Open it in QuantFit (or `qf.load_spec("model.qfspec")` in Python) and you get her panel structure, lags, instrument list, SE flavour, and seed — bit-for-bit.
How do I cite QuantFit in a paper?: Each run produces a citation block in the report builder — APA, Chicago, and BibTeX — including the version, build, and reproducibility hash. The hash resolves to a permanent page on quantfit.app/cite/<hash>.
Does the iOS app run estimators locally, or stream from the desktop?: Locally on Apple Silicon iPads (M1+). On older devices or iPhones it streams from your paired macOS workspace; the spec is sent, the data never leaves your Mac.
I lost my workspace. Can you recover it?: If you’re on Researcher or Lab, yes — workspaces are versioned and we keep 30 days of snapshots. Email support with the workspace ID (in the URL) and we’ll restore it.
Will QuantFit support my estimator (NARDL / FMOLS / DOLS / etc.)?: The 2026 roadmap is public. NARDL and FMOLS are landing in 2026.2; DOLS, panel quantile, and MIDAS are in design. Open a request on the forum and we’ll bump the priority if a research group is waiting on it.
I’m a referee — can I re-run the author’s analysis without buying a license?: Yes. Replication bundles open in our free Reviewer mode — read-only, full output, no fit modification. The author shares a single `.qfpack` file; you drag it onto the app.
How do I get a site license for my department?: Talk to us via the For labs page. Typical turnaround is two weeks from intro to signed order; named users, invoice billing, and data-residency options on request.

System status & further reading

Self-serve, before you email.

Status page

Real-time uptime for the web workspace, sync, and report builder.

Documentation

The manual — every estimator, with assumptions, diagnostics, and references.

Changelog

Every release, every estimator added or refined, with the JEL-style notes.

Help, writtenby people who run regressions.