Frequently Asked Questions¶

Visit the AIVA sign-up page and register using your email address. After signing up, verify your email address by clicking the link sent to your inbox. See Account Setup for a full walkthrough.

AIVA offers five tiers: Free, Trial, Plus, Pro, and Enterprise. Each tier provides progressively more features, higher upload limits, and additional AI capabilities. See Subscription Tiers for a detailed comparison.

Click the tier indicator in the top-right corner of the navigation bar, select Manage Subscription, select tier by clicking Upgrade, and follow the secure payment flow. Your new tier is activated immediately upon successful payment.

Yes. The Trial tier gives you temporary access to the complete feature set. Signing up with a new email address will automatically grant you a trial.

Your data is never deleted when you change tiers. All previously uploaded samples, conversations, and analysis results remain accessible. However, data is only retained for 30 days when downgrading to the Free tier.

All payment processing is handled by Stripe, a PCI-DSS Level 1 certified payment processor. AIVA does not store your credit card information.

For teams larger than 10 users or organizations with custom requirements, contact us for enterprise pricing options. Reach out to tarun@mamidi.ai.

AIVA supports the following formats:

• FASTQ (.fastq, .fastq.gz): FASTQ format, including gzip-compressed files.
• VCF (.vcf, .vcf.gz): Variant Call Format, including gzip-compressed files.
• CSV (.csv): Comma-separated values with a header row.
• TSV (.tsv, .txt): Tab-separated values with a header row.

See Uploading Files for detailed format requirements.

File size and upload count limits depend on your subscription tier. Free accounts cannot upload files, while Plus and Pro accounts have limits on file count and size. Check your current usage via the usage indicator in the header, or see Subscription Tiers.

Yes. Plus, Pro, Trial, and Enterprise accounts can import files directly from Google Cloud Storage (gs://), Amazon S3 (s3://), Azure Blob Storage (az://), or public HTTPS URLs. See Cloud URL Imports.

Small Variant Annotation uses AIVA's annotation engine to annotate variants with multiple databases. When enabled during upload, it adds columns such as gene symbol, consequence type, SIFT and Ditto predictions, and more. It is available on Trial, Plus, Pro, and Enterprise tiers. See Small Variant Annotation.

Limited. You can upload files without Small Variant Annotation or Structural Variant Annotation and explore the raw data immediately. On-demand annotation only supports handful of variants at a time, use the AIVA Chat Variant Annotation tool, which performs real-time lookups against ClinVar, gnomAD, CADD, SIFT, and PolyPhen.

Open the Job Manager and review the error message on the failed job. Common causes include unsupported file formats, malformed headers (VCF files must include the #CHROM line), exceeding your tier's upload limits, or network interruptions during transfer. Fix the underlying issue and re-upload. Please reach out to tarun@mamidi.ai for more assistance.

AIVA processes one file upload at a time. You can submit multiple uploads in sequence, and each will be queued and processed independently. Monitor all jobs in the Job Manager.

AIVA (AI assisted Variant Analysis) is our AI Clinical Analyst. It can query your uploaded samples, search biomedical literature, annotate variants, generate charts, explore gene-drug interactions, and more, all through natural language conversation. See AIVA Chat for a full overview.

AIVA has access to 10+ specialized tools: Genomic Data Query, Web Search, Variant Annotation (ClinVar, gnomAD, CADD, SIFT, PolyPhen), Biomedical Literature search, Code Interpreter, Knowledge Graph, Clinical Trials search, Phenotype-Gene Prioritization, MCP custom integrations, and Task Manager. See AI Tools for details on each.

Yes. You can enable or disable individual tools from the tool configuration panel in the chat input area. This lets you restrict AIVA to specific capabilities or reduce response latency.

Yes. AIVA offers a selection of language models with different tradeoffs between speed, cost, and reasoning capability. See Model Selection.

No. AIVA has read-only access to your sample data. It can query, analyze, and visualize your data but cannot insert, update, or delete rows.

No. AIVA queries are scoped to samples you own or that have been shared with you through a project. It cannot access data belonging to other accounts.

Very long conversations may exceed the language model's context window, causing it to lose track of earlier messages. Either switch to a model with 1M token capacity (e.g., Gemini PRO models) or start a new conversation and re-establish the key context in your first message. Keeping conversations focused on a single topic also helps maintain coherence.

Uploaded files are processed and stored on AIVA's secure cloud infrastructure. Parsed variant data is stored in a secure database. RLS (Row Level Security) is enabled to ensure data isolation.

Only you can access your data by default. If you create a project and invite collaborators, they gain access to the samples assigned to that project based on their assigned role. AIVA staff do not access your data except as required for technical support with your explicit consent.

AIVA includes features designed for HIPAA compliance, including PHI detection and access controls. See Compliance for details on the security measures in place.

Yes. You can delete samples and projects from the samples and projects tabs. Deleted samples and their associated data (flags, comments, etc.) are permanently removed from the database.

When AIVA uses tools like Web Search, Biomedical Literature, or Clinical Trials, it sends search queries to external APIs, but it does not send your raw variant data. The Genomic Data Query tool runs queries locally against our database. Variant annotation lookups send individual variant identifiers (chromosome, position, alleles) to annotation services for lookup but do not transmit your full dataset.

The Free tier includes: limited AI chat queries per day, limited API access, and read-only access to the Playbooks marketplace. See Subscription Tiers for the full comparison.

Plus adds: increased upload and query limits, clinical reports with AI auto-fill, API access, and advanced AI capabilities. See Subscription Tiers.

Pro increases all usage caps (uploads, AI queries) and includes priority support. All Plus features are included. Pro is designed for small to medium-sized teams.

The Model Context Protocol (MCP) allows you to connect AIVA to external tools and services that you configure. This enables custom workflows tailored to your specific bioinformatics pipeline. MCP is available on Trial, Plus, and Pro tiers. See MCP Integration.

AIVA's data table uses virtualized rendering and server-side pagination to handle datasets with millions of rows. If you experience slowness:

• Check your internet connection. Data is fetched page by page from the server.
• Try reducing the number of visible columns using the column visibility controls.
• Close other browser tabs to free up memory.
• If the issue persists, contact support with your browser version and dataset size.

Possible causes:

• High server load: During peak usage, responses may take longer. Wait a moment and try again.
• Complex query: Some queries (e.g., aggregations over millions of rows) require more processing time. AIVA will stream the response as it becomes available.
• Network issues: SSE streaming requires a stable connection. If your connection is intermittent, responses may appear to stall. Refresh the page and try again.

Your subscription tier determines how many uploads and AI queries you can perform. Check your current usage via the usage indicator in the header. To continue, either wait for your limits to reset or upgrade your plan.

AIVA generates charts using matplotlib in a sandboxed Python environment. If a chart fails to render:

• The error message in the chat will indicate what went wrong.
• Try rephrasing your request with more specific instructions (e.g., "Create a bar chart with matplotlib showing variant counts per chromosome").
• If the issue persists, ask AIVA to provide the data as a table instead.

Use the search bar in the conversation sidebar to filter by title. If you cannot find the conversation, it may have been deleted. Deleted conversations cannot be recovered. Consider renaming important conversations with descriptive titles to make them easier to find in the future. You can also export the conversation as a html page.

Small Variant Annotation processes each variant through the annotation pipeline. For large files (whole-genome VCFs with millions of variants), this can take a significant amount of time. The job continues processing in the background. Check the Jobs tab for the current status. If the job appears stuck in the same state for an extended period, it may have encountered an error. Contact support if needed.

Use the Help modal in the application header to access our slack channel. Include details about the issue: what you were trying to do, what happened, any error messages, and your browser/OS information.