How much does ScienceMachine cost?

No public pricing. ScienceMachine offers a free workflow audit and a pilot on real day to day work before a wider rollout.

What are the best alternatives to ScienceMachine?

Based on documented capability coverage across the Agentic Index 14 point taxonomy, the closest alternatives to ScienceMachine among data analyst agents are Potato, Lumi AI, Pluvo, Prophet. Each is graded on the same public evidence, and no vendor pays for placement.

Does ScienceMachine give you tracing and observability into agent runs?

Full under the Agentic Index verification standard. ScienceMachine publishes public evidence of tracing, observability and audit logging into agent runs, so a team can see what an agent did and why. Evidence basis: Every source, automated action, edit, version and approval is captured in an audit trail, reports keep versions, code traces, source links, edits, signatures and exportable audit logs, and Peak Review keeps each source chromatogram with its. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Does ScienceMachine require human approval before an agent acts?

Full under the Agentic Index verification standard. ScienceMachine publishes public evidence of human approval steps and guardrails, so an agent action can be held for review before it executes. Evidence basis: Reporting runs tracked changes and an independent QC step and captures electronic approval before the final record is exported, uncertain peaks are routed to a scientist for the final decision, and low confidence document readings are surfa. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Does ScienceMachine have SOC 2, SSO and role based access control?

Partial under the Agentic Index verification standard. ScienceMachine publishes some public evidence of security and identity controls, but narrower than the full set of SOC 2, single sign on (SSO) and role based access control (RBAC). Evidence basis: ScienceMachine states GDPR, GxP and 21 CFR Part 11 compliance and that customer data is encrypted, isolated and never used to train shared AI models. No independent attestation such as SOC 2 or ISO 27001, and no SSO or role controls, are do. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Back to vendors

ScienceMachine

Q: Who is ScienceMachine for?

Contract research organizations running bioanalysis, Biopharma service providers, Regulated bioanalytical laboratories

Also known as: Science Machine, Science Machine Inc

Visit site

Data analyst agentindependentVerified 2026-09-28

Entry priceCustom (contact sales)Full pricing detail

AI automation for bioanalytical labs at CROs and biopharma service providers, covering LC-MS peak review, study document QC and regulatory report drafting, with scientist approval and a full audit trail.

ScienceMachine sells AI automation to bioanalytical laboratories at contract research organizations and biopharma service providers. Its three workflows cover the review and reporting work around a study. Peak Review collects chromatograms, method context and batch metadata from the lab's LC-MS instrument workflow, checks each peak's integration boundaries, noise and shape with visual models, documents the consistent results and routes uncertain or non conforming ones to a scientist.

Document QC reconciles a study's draft report, tables, calculations and values against its approved sources, protocols, SOPs and QC rules, including scanned and handwritten records, and returns a prioritized list of findings linked to their sources. Reporting fills the client's approved Word template from LIMS output, tables and instrument files, then runs tracked changes, an independent QC step and electronic approval before the final record is exported.

ScienceMachine configures each workflow around the customer's systems, decision rules and approval model and validates the version the team will use. The company states GDPR, GxP and 21 CFR Part 11 compliance, keeps customer data encrypted and isolated and does not use it to train shared models, and records every source, automated action, edit, version and approval in an audit trail. It has offices in London and San Francisco. Its earlier product, an autonomous bioinformatics agent named Sam, is no longer offered on the site.

Vendor details

Canonical URL

https://sciencemachine.ai

Funding status

Pre-seed round of 3.5 million dollars (reported as 2.9 million euros), announced July 2025, led by Revent and Nucleus Capital with participation from Juniper and strategic angels, per Tech.eu, EU-Startups, and the Nucleus announcement. Founded 2024 to 2025 by Lorenzo Sani (CEO) and Benjamin Tenmann, London based with a San Francisco office and a Newark, Delaware legal entity; ran on a team of two at the raise and was already used by biotech customers.

Company status

independent

Use cases & customers

Primary use cases

LC-MS chromatogram peak review with exception routingQuality control of bioanalytical study documentsDrafting regulatory study reports from approved templatesAudit ready records for GxP bioanalysis

Target customers

Contract research organizations running bioanalysisBiopharma service providersRegulated bioanalytical laboratories

Deployment options

SaaS

Integrations

Reads data from Thermo Fisher, SCIEX, Waters, Bruker, Agilent and Shimadzu instrument workflows, Watson LIMS and other LIMS exports, SharePoint folders, spreadsheets, RTF, PDF, Word and scanned documents; exports reports as Word, Excel, PDF, HTML or PowerPoint.

In practice

A CRO's scientists spend days reviewing LC-MS chromatograms peak by peak. Peak Review inspects every peak against the lab's review criteria and sends only the uncertain or non conforming ones to a scientist.

A study report is due to a sponsor and every value has to match its source. Document QC checks the draft, tables and calculations against the protocol, SOPs and instrument files, and links each finding to the exact place in the source.

A lab writes the same kind of regulatory report for many clients. Reporting drafts it in each client's own template, runs an independent QC pass and captures electronic approval before export, with a full audit trail kept with the report.

Sources & related URLs

Research sources

https://tech.eu/2025/07/09/sciencemachine-raises-35m-for-leading-autonomous-ai-in-biotech-research/ https://www.eu-startups.com/2025/07/london-based-sciencemachine-raises-e2-9-million-for-biotech-analysis-with-a-team-of-only-two-and-without-spending-a-cent-on-marketing/ https://medium.com/@nucleuscapital/why-we-invested-in-sciencemachine-f88b1d01ddcb

Agentic Index coverage score

6.5 / 14 capabilities · 46%

Integrations & Tool Calling	Partial
ScienceMachine reads chromatograms and batch metadata from Thermo Fisher, SCIEX, Waters, Bruker, Agilent and Shimadzu instrument workflows, Watson LIMS exports, SharePoint folders and office files, and exports reports as Word, Excel, PDF, HTML or PowerPoint. Reporting also takes content from LIMS output, spreadsheets, RTF files, protocols and study data. Writing results back into a LIMS or other system of record is not described. SourceScienceMachine, sciencemachine.ai/solutions/lc-ms-review and /reportingread 2026-10-06
Workflow Orchestration	Partial
Each solution runs a fixed three step pipeline. Peak Review ingests the batch, inspects every peak and routes exceptions; Document QC takes the study record, checks every detail and returns findings to resolve; Reporting collects the study record, populates the report and moves it through review, QC and signature. ScienceMachine configures each pipeline around the customer's systems, source formats and decision rules, and a flow the buyer builds or edits is not described. SourceScienceMachine, sciencemachine.ai/solutions/lc-ms-review, /document-qc and /reportingread 2026-10-06
Knowledge Grounding & RAG	Full
Document QC reconciles a study's draft report, tables, calculations and values against its approved sources, protocols, SOPs and QC rules, finding missing or incorrect files as well as isolated values, ranges, conclusions and table mismatches. Reporting drafts sections from the protocol, SOP context and the client's own templates and terminology. The inputs include Watson LIMS exports, PDFs, Word and Excel files and scanned and handwritten records. SourceScienceMachine, sciencemachine.ai/solutions/document-qc and /reportingread 2026-10-06
Human Oversight & Guardrails	Full
Reporting runs tracked changes and an independent QC step and captures electronic approval before the final record is exported, uncertain peaks are routed to a scientist for the final decision, and low confidence document readings are surfaced for confirmation. Peak Review also blocks the workflow when required inputs are missing or confidence is insufficient. SourceScienceMachine, sciencemachine.ai/solutions/reporting, /lc-ms-review and /document-qcread 2026-10-06
Security, Identity & Governance	Partial
ScienceMachine states GDPR, GxP and 21 CFR Part 11 compliance and that customer data is encrypted, isolated and never used to train shared AI models. No independent attestation such as SOC 2 or ISO 27001, and no SSO or role controls, are documented. SourceScienceMachine, sciencemachine.airead 2026-10-06
Observability & Auditability	Full
Every source, automated action, edit, version and approval is captured in an audit trail, reports keep versions, code traces, source links, edits, signatures and exportable audit logs, and Peak Review keeps each source chromatogram with its automated assessment and the scientist's decision. Every Document QC finding points back to the source file and location that supports it. SourceScienceMachine, sciencemachine.ai/solutions/reporting, /document-qc and sciencemachine.airead 2026-10-06
Memory & State Persistence	Not documented
No memory the system keeps across batches or studies is described, and no learning from reviewer corrections. What persists between runs is the versioned templates and the audit trail. SourceScienceMachine, sciencemachine.airead 2026-10-06
Deployment & Data Residency	Not documented
No hosting region, private cloud or on premises option is described, and the service runs as a web app at app.sciencemachine.ai. It integrates with the lab's existing tools rather than replacing them. SourceScienceMachine, sciencemachine.airead 2026-10-06
Prebuilt Agents / Templates / Packs	Full
ScienceMachine sells three named, ready made workflows for bioanalytical labs, CROs and biopharma service providers: Peak Review for LC-MS chromatograms, Document QC for study records and Reporting for regulatory report drafts. Each is configured to the customer's SOPs and templates. SourceScienceMachine, sciencemachine.airead 2026-10-06
Triggers & Channel Coverage	Partial
Work starts from a batch collected from the connected instrument workflow or a study record a person provides, inside the web app; no schedule, event trigger or channel beyond the app is described. SourceScienceMachine, sciencemachine.ai/solutions/lc-ms-review and /document-qcread 2026-10-06
Model Flexibility & Routing	Not documented
Peak Review uses visual models to check integration boundaries, noise, peak shape and acquisition anomalies against approved review criteria. No model provider, model choice or routing a buyer can set is named, and customer data is never used to train shared AI models. SourceScienceMachine, sciencemachine.ai/solutions/lc-ms-review and sciencemachine.airead 2026-10-06
APIs / SDKs / MCP Extensibility	Not documented
No API, SDK, MCP server or developer documentation is published; the site links only the web app login. Configuration around each lab's systems is done by ScienceMachine itself. SourceScienceMachine, sciencemachine.airead 2026-10-06
Testing, Debugging & Optimization	Partial
ScienceMachine validates the configured workflow version before a team uses it, and a proposed change to a locked template creates a new version for review and revalidation; no test set, accuracy scoring or method for that validation is described. It reports up to 80% less manual QC review time with Peak Review and up to 80% faster drafting with Reporting, and its customer Inoviv says it cut the time spent on QC and reporting by 80%. SourceScienceMachine, sciencemachine.ai/solutions/lc-ms-review, /reporting and /document-qcread 2026-10-06
Browser / Computer-use	Not documented
No browser or desktop operation is described. ScienceMachine reads scanned and handwritten documents as inputs, and surfaces ambiguous handwriting for a person to confirm. SourceScienceMachine, sciencemachine.ai/solutions/document-qc; sciencemachine.airead 2026-10-06

The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Pricing

Custom (contact sales)

Cost watchouts

Each workflow is configured around the lab's systems, decision rules and approval model and validated before use; whether that setup is charged separately is not stated.

Variable cost rationale

No billing unit is published, so exposure is unknown; volume of batches or studies is a plausible axis for this kind of service, which is an inference.

Sales call required

Yes, required for paid access

Free / trial

Free workflow audit and a pilot; no self serve trial

Key ambiguities

No price, plan or billing unit is published; the site offers a free workflow audit and a pilot before a wider rollout.

Agentic Index verified 2026-09-28

Official pricing page

Related vendors

Alternatives to ScienceMachine

The closest documented capability profiles to ScienceMachine among data analyst agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.

Potato6.0 / 14Adds documented Memory & State Persistence and Deployment & Data Residency
Lumi AI7.5 / 14Adds documented Memory & State Persistence and Deployment & Data Residency
Pluvo6.5 / 14Adds documented Deployment & Data Residency and APIs, SDKs & MCP Extensibility
Prophet3.5 / 14A lighter documented profile than ScienceMachine
Chord8.0 / 14Adds documented Memory & State Persistence and APIs, SDKs & MCP Extensibility
Actian AI Analyst9.5 / 14Adds documented Memory & State Persistence and Deployment & Data Residency, among others

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Compare ScienceMachine with other vendors →Compare all data analyst agents →See where ScienceMachine ranks →

Data confidence: medium