Short answers to the questions that come up most often.
Rigor turns a one-line brief into a structured, verifiable deliverable — competitive scans, gap analyses, risk registers, implementation prompts, and similar. It classifies what you need, drafts a methodology, runs the steps, and hands back a referenced document.
Write what you need in plain language and submit. Rigor proposes a plan — you can review, edit, or re-classify the task before executing. Planning is always free; you only spend credits on execution.
Yes — the composer surfaces 40+ task templates (Competitive scan, Risk register, Bug diagnosis, and more). Pick one, fill in the variables, and submit.
Thoroughness levels trade depth for cost:
You can switch between levels from the plan screen — the step list and credit estimate update in place.
Yes. The plan is fully editable up until you press Execute: skip individual steps, insert framework steps between any two rows, or override the classified task type if it's wrong.
Each step pauses for your review before the next one starts. You can continue, refine, or stop the run — useful when you want tight oversight or are iterating on prompt phrasing.
File attachments aren't supported yet — for now, paste the source text into your task description or link to public URLs. We're working on a proper attachment flow.
Two layers. A monthly plan covers the engine, dashboard, integrations, and tier-level features (concurrent tasks, history, member seats). Credits cover the AI work itself — every task consumes credits based on methodology level and complexity. The plan screen shows an estimated range before you commit — a Standard task typically costs 15,000–22,000 credits. Credits are deducted per step as the run progresses, so you only pay for work that actually happened.
$1 buys 1,000 credits. Same per-credit price on every tier. Plan calls are always free.
Yes. Every new account gets 20,000 free credits on signup, granted once you verify your email — typically enough for one Standard task end-to-end. Credits never expire and the per-credit price is the same on every tier.
The Free plan is also a tier on its own: 1 member, 1 workspace, 1 concurrent workflow, 2 MCP connections, 1 API key, 7-day history. Upgrade to Pro ($15/mo), Team ($79/mo), or Enterprise ($399/mo) when you need more.
Free — Free — 1 seat, single concurrent task, 7-day history. Best for kicking the tyres.
Pro — Pro — $15/mo. Up to 3 members, 3 concurrent tasks, 90-day history, full step outputs, interactive mode, API + webhooks.
Team — Team — $79/mo. Up to 20 members, 10 concurrent tasks, 10 workspaces, LLM policy, workspace budgets.
Enterprise — Enterprise — $399/mo for up to 50 members, unlimited workspaces and concurrent workflows, audit log export, custom roles. Need a larger deployment? Contact sales.
An optional safety net: when your balance falls below a threshold you set (default 5,000 credits), your card is charged for a reload amount you pick. You set a monthly cap so it can't surprise you.
Toggle it on or off any time from Settings → Billing.
If a run fails or returns a clearly broken deliverable, credits are refunded automatically. For other refund requests, write to us with the task ID and we'll work it out.
No. Task content, intermediate steps, and deliverables are never used to train models or shared with model providers beyond what's needed to run your task.
Tasks and deliverables stay in your library indefinitely so you can re-open them. You can delete any task from the library or set workspace-level auto-deletion (30 / 90 / 365 days) under workspace settings.
Data is stored on encrypted infrastructure on US servers (AES-256 at rest, TLS 1.3 in transit). Enterprise plans can request specific data residency arrangements.
From Settings → Team, invite by email and pick a role: Admin, Member, or Viewer. Invitees get a link that expires in 7 days.
Seat limits per plan: Free is single-seat (you), Pro allows 3, Team 20, Enterprise 50. At-limit workspaces see an upgrade prompt instead of the Invite button.
Personal workspaces are private to you — tasks aren't visible to anyone else. Team workspaces share tasks, folders, and billing across members. Switch between them from the avatar menu in the top right.
Not directly yet. As a workaround, export the deliverable as Markdown and re-create the task in the destination workspace. Cross-workspace move is on our roadmap.
Go to Settings → Account → Authentication and click "Set up two-factor". Scan the QR code with 1Password, Authy, or Google Authenticator, then verify with the 6-digit code. We give you 10 backup codes — save them somewhere safe; they let you sign in if you lose your authenticator app.
On the 2FA challenge screen, click "Use a backup code instead". Each backup code works once. If you're out of codes, email support@plith.ai — identity verification takes 3-5 business days.
Settings → Developer → API keys. Each key is shown only once at creation — copy it then. Keys are scoped to your active workspace and inherit your role's permissions. Authenticate with the x-api-key header.
Configure endpoint URLs at Settings → Developer → Webhooks. We POST a JSON payload signed with HMAC-SHA256 (X-Plith-Signature-256 header) whenever your selected events fire. Respond with any 2xx within 5 seconds. Failures are retried with exponential backoff (1s, 3s, 7s, 15s, 30s). 8 event types are subscribable.
The Model Context Protocol lets AI agents and AI-powered tools call Rigor programmatically. The MCP server exposes 5 tools — rigor_plan, rigor_execute, rigor_status, rigor_approve, rigor_step. Auth is by API key, same as REST.
120 requests/min per org (sliding window). Daily cap on /api/rigor/execute is 1,000/day. Concurrent active workflows are tier-gated: 1 on Free, 3 on Pro, 10 on Team, unlimited on Enterprise. Hits return 429 with a Retry-After header.
Steps occasionally take longer than the estimate when sources are slow to respond. If a step has been running for more than 10× its expected duration, click Abort on the task page and re-run — your credits for completed steps are preserved.
Open the step drawer (click any step row) to see what each step produced. Most quality issues trace back to an ambiguous brief — edit the task and re-plan. If a specific step is the culprit, you can re-run just that step from the step drawer.
If a workflow fails before producing a deliverable, Free plan users receive a full credit refund automatically — check the Usage history in Settings → Billing. If the refund hasn't appeared after 10 minutes, drop us a note with the task ID.
On Pro and above, you keep access to all completed step outputs from the partial run, so credits for those steps are not refunded — open the step drawer to see exactly which step failed and what it produced, then re-run only that step or refine the brief. Reach out if a step misbehaved and we'll review the run.
PDF export is coming soon. For now, export to Markdown and convert with your tool of choice — most editors render it cleanly to print.
Drop us a line at support@plith.ai — include the task ID if it's about a specific run.