Interactive execution canvas
Three-column bench: icon rail, experiment logs, and protocol templates beside the live step stream. Timers and the prompt bar stay on the run.
runline
The runtime execution engine for researchers
Protocols step-by-step, media prep that scales, and triage at the bench. Community is free with Google or GitHub — Individual Pro $18/mo in test checkout.
Unpublished science stays unpublished · No model training on your data
runline.space/notebook
Step 2 · Assemble 20 µL reactions
Scale for 24 wells at +10% overage. Confirm water volume before plating.
Countdown timer
8:00 remaining · 37 °C hold
Countdown timerMedia prepDilution math
An execution assistant, not a filing cabinet
Live today: the three-column bench, pre-run setup, modular tools, protocol library, analyzer, quantify, and diagnostics.
Pre-run protocol setup
Dragging a template onto the canvas pauses for sample count, replicates, plate format, and pipetting overage — volumes are right before step 1.
Dynamic media prep scaling
Change a batch volume and ingredient masses recalc on the card. LB, agar, and master-mix overage stay consistent without a side spreadsheet.
Modular step tools
Every step has + Add Tool. Attach a timer, dilution calculator, or media-prep card only when that run needs it.
Bench diagnostics
Runline knows what you actually did, what your lab has, and what worked before. Capture a failed run, rank causes with source badges, and prepare for your PI — never a mandatory sign-off, never a compliance claim.
Dedicated protocol library
Assay templates live in their own library — not mixed into run logs. Ingest Methods text or figures into an executable DAG.
From a methods paragraph to a lab that remembers
Runline knows what you actually did, what your lab has, and what worked before — ingest, configure, execute, diagnose, export.
Ingest & graph
Paste Materials & Methods or drop a paper screenshot. The compiler parses the method into an executable DAG in the Protocol Library. Figures are read by in-browser OCR, with an AI read added when AI is on.
qPCR · Western · ELISA · Bradford · plaque
Pre-run setup
Configure sample count, replicates, plate format, and pipetting margins before anyone starts the first incubation.
24 samples × 3 replicates · +10% overage
Execute with modular tools
Follow the bench. Add a timer, media-prep card, or dilution calculator on any step. Scale master mix in place.
+ Add Timer · + Add Calculator · media scale
Triage & diagnose
Photograph smeared gels or open Diagnostics: offline analysis always works; paid AI is optional and labelled. Every hypothesis carries a source badge — Rule, Playbook, Your lab, Paper, Expert, or AI hypothesis — unverified.
Offline analysis · source badges · consult PI (optional)
Quantify & export
Assay tables and timestamped parameters write into the run log. PDF / Markdown export is live; Lab HQ shares runs with your team.
Exportable run logs
Templates live in the library — not the log stream
Ingest published Materials & Methods or a paper screenshot, compile an executable DAG, and keep run history clean.
Text ingestion
Paste a Methods paragraph. One click compiles steps, timers, and buffers into a runnable graph.
Figure ingestion
Drop a methods figure. In-browser OCR reads the columns (plus an AI read when AI is on), and the compiler emits the DAG you launch on the bench.
Out-of-the-box assays
Featured starters — open Protocols for the full catalog.
Try the live demo
Visual protocol runner, instrument error decoder, and image triage — local mocks when AI is off. Demo photos stay in browser memory; they are never written to storage.
Ready for a real run? Open the notebook workspace.
Community is free · Individual Pro $18/mo · Lab Pro $15/seat/mo
Use the live Community workspace now. Individual Pro is $18/mo or $180/yr; Lab Pro is $15 per seat per month (minimum 3 seats). Enterprise remains coming soon.
Your unpublished work stays unpublished
Labs do not adopt tools that put a grant at risk. Proprietary protocols and bench observations are not used to train foundation models under our paid AI configuration; see /ai. We make no regulated-system validation claims.
Unpublished science protection
When AI features are on, prompts go to Anthropic Claude by default, with Google Gemini as fallback and for audio and image reading. Both process requests under commercial API terms (see /ai). When AI is off, nothing is sent to either provider. Inference is ephemeral; providers may retain encrypted abuse metadata briefly — not a Zero Data Retention claim.
- · We do not train models on your content
- · Commercial API terms with every provider
Encryption in transit and at rest
Encrypted in transit and at rest by our hosting providers. Lab-scoped tables use Row-Level Security and RBAC. The shared reagent catalog holds vendor catalog metadata (for example NEB, Thermo Fisher, ATCC), not lab data; signed-in users can read it but not write to it. It is shared by design, so we do not claim the catalog is tenant-isolated.
- · Encrypted in transit and at rest by our hosting providers
- · Lab-scoped RLS · shared vendor catalog is read-only (catalog is not isolated)
Academic integrity guardrails
runline critiques and verifies methodology — it does not ghostwrite aims or invent experimental results. Diagnostics never labels a person with integrity accusations; those checks live only under /review.
- · Claims carry sources
- · Deviation and CAPA log
AI accuracy & calculation verification
Justin and generative surfaces are assistive aids. The executing scientist verifies volumes, containment, and chemical safety before pipetting.
- · Verify buffer math before execution
- · Scientist remains accountable
Coming SoonSOC 2 Type II — coming soon. University SAML SSO is not shipped. DURC / dual-use screening is not productized. Draft DPA and security one-pager: download /dpa. Readiness stub: /soc2.
Privacy inquiries: support@runline.space
Tell us about the lab
Instruments, cohort size, and purchasing path — we onboard labs in small cohorts so every protocol library gets indexed properly.
Submissions go to support@runline.space.