Know whether your pipeline is still validated after every change.
Aver benchmarks your clinical pipeline against public reference materials, detects when a change moves the numbers, and produces signed evidence your inspector can verify.
Live product demo
Your pipeline changes even when nobody touches it.
Registries rebuild images, a colleague bumps a tool, an annotation cache updates. Any of these can move variant calls. Manual revalidation happens in batches, so the gap stays invisible until an inspection finds it. Aver records every change as it happens, with its measured effect.
| Detected | What changed | Severity | Measured impact |
|---|---|---|---|
| Day 1 | bwa-mem2 container rebuilt upstream Digest changed; every tool version inside identical | Documentation | Not re-run |
| Day 14 | samtools 1.21 → 1.22 Minor version bump; benchmark re-run automatically | Review required | SNV recall ±0.00 pp |
| Day 29 | GATK 4.5.0.0 → 4.6.0.0 SNV recall fell beyond the −0.5 pp threshold | Revalidation required | SNV recall −0.87 pp |
| Day 41 | multiqc pinned to :latestMutable tag in a pipeline marked clinical | Critical | Approvers alerted |
Five steps, then it runs in the background.
Revalidation stops being a project you schedule and becomes a record that keeps itself current.
- 01
Connect
Read-only access to your Nextflow, Snakemake or WDL repository, or a container and parameter file.
- 02
Fingerprint
Every tool, version, container digest, parameter, genome build and annotation release, resolved and hashed.
- 03
Benchmark
Run against GIAB reference materials and compare with hap.py and vcfeval. Precision, recall and F-measure per variant class.
- 04
Watch drift
When anything changes, Aver re-runs the benchmark and measures the effect against your approved baseline.
- 05
Evidence pack
A timestamped, signed report mapped to checklist requirements, ready for your quality system.
Severity comes from rules you can read.
No model decides whether a change matters. Each flag comes from an explicit, versioned rule, so the reasoning holds up when someone asks why.
Evidence an inspector can check, not just read.
Designed for laboratories working under CAP, CLIA, ISO 15189 or IVDR expectations. Aver generates the record; your laboratory makes the decisions.
- DeterministicRegenerating the same period produces a byte-identical document.
- Signed with re-authenticationIdentity, role, meaning and time are bound to one exact version.
- Hash-chained audit historyEvery baseline, change and approval carries the hash of the entry before it. Any edit breaks the chain.
- Verifiable by anyoneVerify recomputes the document from the underlying records and confirms the hashes match.
Change impact summary
| Change | Impact | Disposition |
|---|---|---|
| samtools 1.21 → 1.22 | ±0.00 pp | Acknowledged |
| bwa-mem2 digest rebuilt | Not re-run | Recorded |
| Scheduled verification | +0.01 pp | Within baseline |
sha256:3a667de0f…
Sits beside your lab stack, not in place of it.
Keep your workflow engine, LIMS and quality system. Aver adds the validation layer between them.
No patient data, ever
There is no field anywhere for a patient identifier, clinical sample or annotation. Benchmarks use public reference materials.
Runs where your data lives
The lab Runner executes inside your environment with Docker or Apptainer, connects outbound on port 443, and returns metrics and metadata only.
Works with the engines you use
Dependency resolution for the major workflow managers, or register a container and parameters manually.
Your record stays yours
Evidence packs and audit history remain downloadable even if you cancel. They are your regulatory record.
Built for teams that must prove their analysis is right.
Clinical genomics labs
Laboratories operating under CLIA, CAP or ISO 15189, running NGS pipelines that need revalidation on every change.
IVD and diagnostics
Teams pursuing or maintaining IVDR conformity, where the pipeline is regulated as medical device software.
Pharma and CRO informatics
Computational workflows operating under GxP expectations for qualification and change control.
Research bioinformatics
Groups who want reproducibility evidence and benchmark tracking without a regulatory obligation.
See a real number before you pay anything.
Start on the free CLI. Move to the platform when you need continuous monitoring and evidence on demand.
Open source CLI
For evaluating Aver, research work, or trying it before involving anyone else.
- Local benchmarking and reports
- Dependency fingerprinting
- Snapshot diff
Lab
For laboratories running accredited clinical testing that need evidence on demand.
- Everything in the CLI
- Continuous monitoring, drift detection and alerts
- Hosted history and evidence packs
- Up to three pipelines
Enterprise
For teams under IVDR or GxP, or with signature workflows inside a quality system.
- Everything in Lab
- Unlimited pipelines
- Electronic signatures and checklist mapping
- Audit log export, SSO, priority support
Annual billing is available and is the norm for laboratory procurement. Purchase orders are supported on Enterprise.
Run your first benchmark locally.
Fully local. No account, and no network calls beyond downloading reference data. Push results to the platform later if you want continuous monitoring.
# install and run locally pip install aver-cli cd /path/to/your/pipeline aver init aver resolve aver benchmark --sample HG002 --build GRCh38