AIxBioinformatics
CN
✕
CN EN
What the platform does

You decide the analysis; we get the compute done

Placement that follows your constraints

Each job is placed according to data sensitivity, memory and GPU needs, deadline and budget, with the reasons for and against each option written down. If a reclaimable resource is withdrawn, the job is queued again; after two withdrawals it moves to resources that are not reclaimed.

Ready-made standard pipelines

Eleven community-standard pipelines covering variant calling, RNA-seq, single-cell, methylation, ATAC and ChIP, metagenomics, amplicon and protein structure prediction, at fixed versions with traceable sources.

Built for AI from the start

Describe the job in a sentence and AI turns it into a compute plan for you to confirm. A REST API and an MCP endpoint let your own AI assistant list pipelines, check placement, submit requests and follow progress.

Apply for access
How it works

People check each step; nothing runs before you accept the quote

1

Submit

Describe it on the web, or let your AI assistant submit through the API.

2

Review and quote

We check the data type and the run plan by hand and send a written quote.

3

Accept

Nothing is scheduled until you accept; no charge if you do not.

4

Check data, schedule

We confirm where the data is, set the run parameters and hand it to the scheduler.

5

Run and deliver

Progress, placement and the processing log are visible in your console; billing happens only when the run completes.

Price match: For compute resources of the same configuration and duration, if you get a lower written quote, we will match that price.

Data and compliance

Human data runs only on our compliant local cluster

  • Anything involving human samples is treated as human genetic resources and runs only on the local compliant cluster, never on elastic resources.
  • Data type is checked by hand before quoting, not taken on trust.
  • The scheduler holds orders and run records only; raw data never passes through it. These order and account records are currently stored on servers outside mainland China and will move to mainland China before full launch.
  • Plain-language requests are structured by an AI service hosted outside mainland China; please do not include personal or medical details.
  • Every placement, reassignment and manual action is logged.
Which to choose

Compute versus bioinformatics services

  • Research Compute: you choose the pipeline and samples, we agree the run settings with you, run it and deliver the result files and run record.
  • Bioinformatics services: we do the analysis and experts review it before delivering the report, figures and reproducible code.
  • Not sure? Ask us. One project can use both: compute first, then analysis.
Bioinformatics services
FAQ

About Research Compute

Can I use it now?
The platform is opening in stages. You can apply now, and we will reply within one working day to confirm the available resources, the quote and when you can start. Once access is set up, we explain how to get started.
Which pipelines are available?
Eleven at present: sarek, raredisease, rnaseq, scrnaseq, methylseq, atacseq, chipseq, mag, taxprofiler, ampliseq, proteinfold. All are community-maintained open-source pipelines at fixed versions. Tell us if you need another one.
How does my AI assistant connect?
Once you have access, create an API key in the console and follow the developer docs for the REST API or MCP. Accepting a quote needs a key with payment permission switched on, and the account holder's consent first.
How is it priced?
Each job is quoted. It runs only after you accept the written quote, and is billed only when the run completes; failed or cancelled jobs are not billed. You can cancel yourself before accepting the quote; after that, contact us. For compute resources of the same configuration and duration, if you get a lower written quote, we will match that price.
How long does a run take?
It depends on the pipeline, sample count and available resources. Once each pipeline has been benchmarked the console shows an expected finish time; until then we estimate per project.
Is the platform in English?
The interface is currently in Chinese and serves research teams in mainland China.

Apply for Research Compute

Tell us the pipeline, sample count and where your data sits; we reply within one working day to confirm the available resources, the quote and when you can start.

Apply for access