Studio Limits Reference

Limits and Quotas at a Glance

Every number Studio enforces, in one place: plan limits, per-feature caps, sizes, timeouts and retention. The pages linked in each section explain what happens when you reach a limit and what to do about it.

Plans are per organization. A limit you are already over — after a downgrade or when a trial ends — never locks existing work: you can keep saving and deleting, you just cannot add more.

---

Plan limits

	Free	Plus	Pro	Corporate
Price shown in the app	Free	$25/mo	$99/mo	Custom (Contact sales)
Projects per organization	2	10	unlimited	unlimited
Members per organization (owner and pending invitations count; an invitation sent again counts once)	2	15	unlimited	unlimited
Completed interviews per project (all environments together; screen-outs, partials and the Example template's sample rows not counted)	1,000	unlimited	unlimited	unlimited
File storage per organization (uploads and run outputs)	250 MB	5 GB	50 GB	unlimited
Analysis flows per project (checked when a Save adds one: a project started from the example study keeps its six on every plan)	3	20	unlimited	unlimited
Access codes per questionnaire	100	5,000	unlimited	unlimited
One flow run: time / memory	5 min / 512 MB	15 min / 1 GB	30 min / 2 GB	30 min / 2 GB
"Run to here" previews per user, per project, per hour	30	120	600	unlimited
Previews running at once per user	1	1	2	4
Custom JavaScript and custom CSS in a published survey	—	✓	✓	✓
Live auto-recompute and public live links	—	✓	✓	✓
Schedules · webhooks · connectors	—	✓	✓	✓
Connectors available	—	Sheets, Excel 365, Supabase, HubSpot	+ S3/R2/MinIO, GCS, Azure Blob, BigQuery, Snowflake, Postgres, SFTP, REDCap, Salesforce, HTTP	+ (coming) MCP
Saving to the organization library	—	✓	✓	✓
Email invitations: per month / per day / first mailing	—	1,000 / 300 / 200	5,000 / 1,500 / 500	no caps
AI assistant: credits per month / per day	—	8,000 / 2,000	50,000 / 8,000	300,000 / 30,000
AI requests per person per hour	—	30	100	300
Larger AI model (drafts from a brief)	—	—	✓	✓
SSO	—	—	after the beta	after the beta
Self-hosting	—	—	—	✓

Free on every plan: the whole Builder (all question types, logic, quotas, randomization, theme, the script library), imports, the question bank and templates, coding open answers by hand and by rules, Download .py, research bundles, every data export, History, pre-registration and deposits, comments and edit locks, API keys.

→ Plans, Trial and Billing

The trial

Length	30 days of Pro, once per email address
Not included during the trial	email invitations (import and send unlock once a plan is bought)
AI during the trial	500 credits in total, at most 200 per day and 10 requests per hour
Reminders	email 7 days and 1 day before the end; banner in the last 3 days
When it ends	the organization moves to the Free plan, or to Plus if it was bought for after the trial; nothing is deleted
Extra organizations you create	start on Free, no trial

---

Accounts and organizations

Item	Limit
Password	at least 8 characters, with a lower-case letter, an upper-case letter, a number and a symbol
Name	1–200 characters; the field stops at 200
Organization name	1–120 characters, on one line; the field stops at 120
Organization slug	3–40 characters (a–z, 0–9, -), made from the name: for an organization you create, as a project slug is (other scripts spelled in Latin letters, organization when nothing can be spelled, -org added to a slug under 3 characters or of exactly 12 hex digits, -2, -3, … when any organization on Studio already has it, checked with the server before the dialog shows it); for the one you get at sign-up, your name with -org added; permanent, unique across Studio
Team invitation link	valid 7 days
API key name	1–80 characters (the field stops at 80); the key is shown once; keys made in the app do not expire
Email lookup on the sign-in page	10 per minute per address and network, and 20 per minute per network across all addresses
Organization house style (Branding)	Custom CSS up to 64 KB; every other value up to 4 KB; 128 KB for the whole style

Projects and the Builder

Item	Limit
Project slug	3–64 characters (a–z, 0–9, -), made from the name — other scripts spelled in Latin letters, a long name cut at a word break, project when nothing can be spelled, -project added to a name of one or two characters (Q1 → q1-project), -2, -3, … when the organization already has it; permanent, unique in the organization
Project name (rename)	1–120 characters
Default environments of a new project	pilot capped at 50 completed interviews, main capped at 1,200
Undo history	100 steps; cleared by each Save
Draft autosave	1.5 seconds after your last change
Save message	up to 240 characters
Likert points	2–11
Custom JavaScript per script	2,000 characters
Import file	up to 2 MB, UTF-8
Library item	name up to 120 characters, description up to 500
Share-preview links	valid 24 hours; up to 50 new links per person per day
Simulate	1–5,000 responses (default 200, seed 42); preview shows the first 50 rows
Edit lock	renewed every 30 s; expires 2 minutes after the last renewal
Pages, questions, options	no fixed limit

Fieldwork

Item	Limit
Survey link	study.siamang.org/<12-character survey id>/, one per environment, stable across republishes
Response cap	the tighter of the environment's cap (counted in that environment) and the plan's (Free: 1,000 per project, all environments together); counts completed interviews only — screen-outs and partials never count, and a screen-out is recorded even when the cap is full; checked when the survey page opens and when a respondent submits
Quota cell	closes at its limit of completed interviews; checked when the respondent leaves a page; a check that gets no answer within 4 seconds lets the respondent go on; previews never check
Closing date	the earlier of the questionnaire's deadline and the environment's closes_at as of the published Save, or a date set in the card's Closing date panel (applies at once); once it passes, the page shows the closed notice as it opens, and submissions and progress saves are refused (a date without a time zone is read as UTC)
One per browser	off by default; one interview per browser per environment, remembered by a mark in the respondent's browser that is not cleared after a week (it stays until they clear their browser data)
Preview deployments	removed automatically after 7 days
URL parameters stored per response	the first 8; names lower-cased, up to 40 characters; values up to 200 characters
Browser autosave for respondents	7 days after the last answer, same browser only, one per survey; nothing is kept before the respondent starts; cleared once the interview is submitted or ended by a full quota. Everything else a survey keeps in the browser also goes after 7 days without a write, except the One per browser mark
Submit retries	3 attempts
Redirect delays	completion 5 s; terminal page 5 s by default; quota-full and full-cap 3 s; closed notice to the environment's post-close redirect 3 s
Access codes	format PREFIX-NNNN (prefix up to 8 characters, 10,000 codes per prefix); 1–5,000 generated at a time
Captcha: completions without a token	3 per hour per survey per network address
Submissions	60 requests per minute per survey per network address; progress saves, the status check as the page opens and quota checks each have a separate allowance of the same size
One submission	answers up to 256 KB and 2,000 answer keys (key names up to 200 characters); the whole request up to 2 MB
Contacts import	up to 20,000 lines per import
Mailing	subject up to 200 characters; message up to 20,000; sent in batches; month and day boundaries in UTC
Automatic mailing pause	complaints ≥ 1 and ≥ 0.05 % of delivered, or bounces ≥ 3 and ≥ 2 %
Live tiles auto-recompute	10 seconds after the last completed response (Plus and above)
Monitor refresh	Distribute and Live every 30 seconds while the tab is visible

Data

Item	Limit
Data grid	100 rows of a table, newest first (tables with a created_at or id column), 25 per page; Search all rows searches every row and loads the newest 100 matches
Export (Data tab, API, bundle with data)	up to 100,000 rows per file
Insights	top 50 values per frequency; top 2,500 cells per crosstab
Panel outcome CSVs	every response of the outcome
Response retention	kept until you delete them or the project

Flows and reports

Item	Limit
Flow run	1 CPU, no internet; time and memory by plan (above)
Outputs collected per run	50 files, 200 MB in total, written under outputs/; when a run writes more, report documents (.md, .html) are kept ahead of figures and other files
Preview run	512 MB, 2 minutes, on every plan
Run history shown	the latest 50 runs
Output download links	valid 5 minutes
Flow and connector names	lower-case letters, digits, _; start with a letter; up to 63 characters; not survey
A node's output File name	Latin letters (A–Z, a–z), digits, -, _ and ., with / between folders; up to 100 characters; always in outputs/
Data file reading	the formats .csv, .tsv, .txt, .xlsx, .xlsm, .xls, .sav, .dta, .parquet; a file's columns are worked out from its first 50,000 rows; a reading that has not answered in 5 minutes is dropped, and read again when a Data file node asks for it
Write table column names	up to 63 bytes each (a Cyrillic letter takes two)
Report custom CSS	anything except the sequence </

Statistics in flows

The methods themselves cap a few computations; each result says which way it went.

Item	Limit
Fisher's exact test, tables larger than 2 × 2	summed exactly over at most 200,000 tables with the observed margins; beyond that, 20,000 random tables from a fixed seed (the same p on every run)
Wilcoxon signed-rank exact p (p-value auto)	up to 50 pairs with no ties or zeros, up to 13 with them; the normal approximation above
Wilcoxon exact p on request (p-value exact)	up to 1,000 pairs; above, the normal approximation, and the result says so
McNemar exact binomial p (p-value auto)	fewer than 25 respondents who answered the two questions differently; the chi-square approximation from 25
Correlation	at least three complete pairs
Factor analysis	at least 3 items and more respondents than items; parallel analysis draws 100 random data sets (95th percentile); maximum likelihood is started from 14 fixed points
Data check	up to 5 example values per problem, then "… (N more)"
TURF exhaustive search (Search best)	up to 200,000 combinations; beyond, the run stops and suggests a smaller portfolio, fewer options or greedy
Key drivers	at least 2 drivers; the Shapley value (Importance shapley) for at most 15 — Johnson's relative weights for more
Significance letters (Banner table, Bar chart, Tab book)	a column or group of fewer than 30 respondents is not tested (the Bar chart and the Tab book count those who answered the question); Level 0.001–0.2

Charts, tab books and report colors

Item	Limit
Figure size	2–30 inches each way (10 × 6 by default); a chart whose labels need more room grows taller
Bar chart, newer forms	a number of more than 30 different values given is refused as bars, a split or a donut (a histogram draws it); with Top N, up to 30 of its values given most are drawn
Bar chart Top N	1–100 answers
Bar chart Bins	at most 100 bins; a number of bins from 1 to 100
Bar chart Confidence	0.5–0.999 (0.95 by default)
Donut Other below (%)	0–50 % (3 by default)
Trend	at most 500 points: the periods from the first date to the last, or the wave codes; bands for up to 4 lines (more lines: marker shapes, the table keeps the intervals); Minimum base 30 by default
Tab book	a banner variable of at most 30 different values; a question without answer labels of at most 30 different answers (a number shows its mean instead); sheet names cut to Excel's 31 characters; the letters as above
Likert chart scale	from value labels, or a valid range of 2 to 11 whole numbers
Chart colors (a report's Look)	Series 2–12 hex colors, none twice; a series or Magnitude color at least 1.3:1 on white; Chart text at least 4.5:1

Coding open answers

Coding by hand and by rules, in the codeframe editor and the Code open answers node. → Coding Open Answers

Item	Limit
AI coding of open answers	switched off on this platform
Codeframe name	analysis/<name>.codeframe.json: lower-case letters, digits, _; starts with a letter; up to 63 characters
Themes per codeframe	up to 200 (an older, version 1 codeframe: 64)
Terms per list (Words and phrases, But not, each Must also contain list)	up to 500
A term	up to 200 characters; words near each other at most ~20 apart; no regular expressions
Replacements	up to 2,000, each side up to 200 characters
Negation	reaches the next 3 words, within the clause; English only
A "word" in an answer	longer than 200 characters (a pasted link), it matches no term
Keys in the answers list	1–9 for the first nine themes
Distinct answers the editor reads	the 100,000 most frequent (beyond: "the rarest answers are left out")
Answers per page in the editor	100
New answers in the editor	read again once the last reading is a minute old
Suggested words	up to 30 words and 30 two-word phrases, each used by 2 respondents or more
Editor requests (answers, previews, tests, suggestions)	60 per 10 seconds per person; one request up to 8 MB
Errors and warnings listed	the first 200 of each, with the total
Credits	none

History, files and automation

Item	Limit
Saves	kept for the life of the project; History lists the latest 100
Comments	up to 4,000 characters; the oldest 500 of a project are loaded
File upload	up to 50 MB per file, refused in the Upload dialog before it is sent; the stored name up to 128 characters, its extension kept; a workbook over 20 MB gets a warning that it reads slowly
File download links	valid 5 minutes
Connector rows	100,000 per run (Google Sheets 50,000; Excel 365 10,000); a larger table fails the whole run
Schedules	5-field cron, UTC, checked every minute
Webhooks	URL up to 500 characters; up to 5 attempts, retried after 2, 4, 8 and 16 minutes; 10-second timeout
Deposits	2-minute timeout per request to Zenodo or OSF
Activity log	latest 100 events shown; export as CSV

See also

• Plans, Trial and Billing
• FAQ and Troubleshooting
• Security and Privacy
• Glossary

---

← Recipes · Studio contents · Keyboard Shortcuts →