Mantis
Mantis is a private MIT CSAIL-developed cognitive cartography platform that turns multimodal data into visual semantic spaces for human-AI collaborative analysis. It serves researchers, data scientists, and knowledge workers via a hybrid self-serve and enterprise go-to-market.
- Company typePrivate
- Founded2025
- HeadquartersCambridge, United States
- Headcount11–50
- GTM typeB2B
- OfferingSoftware
What Mantis does
Mantis is a private cognitive cartography platform developed within MIT CSAIL, providing a visual workbench for multimodal data exploration and human-AI collaborative discovery. The platform ingests diverse data types—including text, PDFs, tables, images, audio, video, molecular coordinates, H5AD single-cell data, GitHub repositories, and Hugging Face datasets—into semantic spaces that users can navigate via interactive maps, lasso and path tools, and category/cluster overlays. Composer and MantisGPT agents operate with map-grounded context to produce narration, summarization, and evidence-linked responses, while Visual Workflows let users compose node-based pipelines with custom Python steps and agent orchestration.
The underlying AI layer is not proprietary: users select between OpenAI chat/embeddings and private, server-hosted vLLM services per space, with no disclosed in-house model training. Mantis operates a hybrid go-to-market combining a self-serve "Open app" product-led path (mantis.csail.mit.edu/home) with a "Request demo" enterprise sales motion (home.withmantis.com/inquire). Pricing is not publicly disclosed but appears subscription-oriented with annual billing and quote-based enterprise terms. The company is headquartered in Cambridge, MA, within MIT CSAIL, affiliated with the Kellis Lab, was founded in 2025, and reports 11–50 employees. No venture funding, named customers, or commercial revenue figures are disclosed in available source material.
Mantis firmographics
Firmographics- Name
- Mantis
- Legal name
- Mantis
- Website
- https://withmantis.com
- Company type
- Private
- Founded year
- 2025
- Operating status
- Operating
- Headcount range
- 11–50 employees
- Short description
- Mantis is a private MIT CSAIL-developed cognitive cartography platform that turns multimodal data into visual semantic spaces for human-AI collaborative analysis. It serves researchers, data scientists, and knowledge workers via a hybrid self-serve and enterprise go-to-market.
- Ownership category
- akta.pro rank
Mantis industry classification
Industry- Product category
- AI Data Analytics Platform
- NAICS
- Software Publishers (5132)
- SIC
- Services-Prepackaged Software (7372), Services-Computer Programming, Data Processing, Etc. (7370)
- akta.pro primary industry
- Enterprise AI Data & Knowledge Platforms (Vector Databases, Knowledge Graphs) (HDAEANAH)
- akta.pro secondary industry
- Generative AI & LLM Solutions Services (RAG, Agents, Copilots) (BPAEAHAG)
Keywords
Where Mantis is headquartered
LocationHeadquarters
- HQ city
- Cambridge
- HQ country
- United States
- HQ region
- North America
Offices1 record
Markets served
Mantis business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Technology or R&D, Personnel, Infrastructure, Operations, Marketing or Sales
Revenue model
- Platform Access and Services: Mantis offers platform access with model and service choices including OpenAI chat and embeddings as well as private vLLM embeddings and instruct services. Users can choose model selection per space, suggesting a consumption or subscription-based model for AI compute and platform usage.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Subscription | Annual | Platform access with model/service options |
Go-to-market motion1 record
Distribution channels2 records
Marketing channels4 records
Mantis product offering
Product offeringCore offering
Mantis is an interactive cognitive cartography platform that transforms multimodal knowledge (text, PDFs, tables, images, audio, video, coordinates, molecules, and biological data) into shared visual maps where people and AI reason together while preserving evidence, provenance, and missing context. The platform combines Semantic Spaces for visual data mapping, Composer and MantisGPT agents grounded in map context, Visual Workflows for node-based reusable analysis, and Developer Tools (MCP, CLI, developer API) for extensibility.
Product overview
Mantis is a unified cognitive cartography platform developed at MIT CSAIL that transforms multimodal knowledge into shared visual maps for human-AI discovery. The platform architecture consists of Semantic Spaces as the core visualization layer for mapping complex multi-source data; Composer and MantisGPT as the agent interface for analysis with map-grounded context; Multimodal Spaces for processing text, PDFs, tables, images, audio, video, and scientific data; Visual Workflows for building node-based reusable analysis pipelines; and Developer Tools including MCP, CLI, and API for extensibility. Users can select between OpenAI or private vLLM models per space. The platform emphasizes visual-first data science where people and AI reason together while maintaining evidence, provenance, and missing context visibility.
Differentiator
Problem solved
Functional benefit
Products and services
- Semantic Spaces Visual spaces that turn raw documents, tables, media, repositories, and scientific data into mappable structures where relationships become visible. Supports multi-source data ingestion, automated embedding and clustering, interactive maps, bags, trees, floors, and lasso/path exploration tools. Targeted at researchers and data scientists.
- Composer and MantisGPT Agent interface that maintains map context visibility throughout analysis. Powers narration, summarization, and evidence-linked responses via MantisGPT, with reusable agent tools grounded in semantic space context. Targeted at knowledge workers and researchers.
- Visual Workflows Node-based data workflow builder for composing reusable workflows, connecting steps, scheduling runs, and carrying mapped evidence into repeatable analysis. Includes custom Python steps, journeys and waypoints, agent orchestration, and scheduled/reusable runs. Targeted at data scientists and research teams.
- Developer Tools Extensibility layer including Mantis MCP (Mantis Cognitive Protocol) tools, CLI for Codex, Claude Code, and OpenCode, custom panels and extensions, interactive notebooks, and Developer API for building integrations and custom applications. Targeted at developers and engineering teams.
- Models and Services Flexible model selection offering OpenAI chat and embeddings as well as private, server-hosted vLLM embeddings and instruct services per space for chat, embedding, and instruction-following tasks. Targeted at enterprise users requiring data sovereignty and AI compute flexibility.
Companies that use Mantis
Customer profileSegments1 record
Ideal customer profiles2 records
Mantis technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration4 records
AI capability10 records
Feature5 records
Mantis partnerships and signals
Strategic signalPartnerships
Three partnerships are on record, tiered core.
- OpenAIcoreIntegration with OpenAI for chat and embeddings services. Users can select OpenAI models per space for AI interactions within Mantis cognitive cartography platform.
- vLLMcorePrivate vLLM embeddings and instruct services available for deployment within Mantis spaces. Provides option for private, server-hosted LLM services alongside OpenAI options.
- MCP (Mantis Cognitive Protocol)coreMCP is Mantis's own protocol for developer tools integration, enabling CLI access for Codex, Claude Code, OpenCode, and extending platform capabilities through custom panels and extensions.
Recent moves6 records
Expansion highlights5 records
Mantis competitors and assessment
Company assessmentEmerging players
- Pinecone: Pinecone is a vector database platform powering semantic search and RAG applications. While it operates at the infrastructure layer rather than the visualization layer, its embedding-based semantic technology underlies the same knowledge organization paradigm Mantis uses.
- Marimo: Marimo is an open-source reactive notebook for Python with reproducibility and Git integration. It is comparable to Mantis in targeting developer-friendly, reproducible data analysis workflows with extensibility through Python.
Direct peers
- Weights & Biases: Weights & Biases provides an ML experimentation, observability, and collaboration platform for data scientists. Its visualization of model runs and datasets is conceptually adjacent to Mantis's visual mapping of multimodal scientific data.
- Observable: Observable is a collaborative notebook and data visualization platform built on JavaScript. It is comparable to Mantis in its visual-first data exploration philosophy and developer-friendly extensibility for knowledge work and analytics.
- Hex: Hex is an AI-powered collaborative data notebook platform that combines SQL, Python, and no-code with AI agents for data exploration. It overlaps directly with Mantis's blend of visual data science, multimodal ingestion, and AI-assisted analysis workflows.
- Noteable: Noteable is a collaborative notebook platform with built-in AI assistance for data science. It shares Mantis's positioning around interactive, AI-augmented analysis of complex multi-source data with visual outputs.
- Deepnote: Deepnote offers AI-powered, collaborative data science notebooks with multi-language support and integrations. It overlaps with Mantis's multimodal data ingestion and AI-assisted analyst workflows for research and data teams.
Broad incumbents
- Palantir Technologies: Palantir's Foundry platform offers enterprise-scale data integration, ontology, and AI-assisted analytics. While far broader and more enterprise-focused, it competes for the same multimodal data exploration and AI-driven decision workflows that Mantis targets.
- DataRobot: DataRobot provides an enterprise AI/ML platform with model building, deployment, and monitoring. Its agentic and multimodal analytics capabilities position it as a broader incumbent targeting similar research-and-data-science buyer personas.
- Tableau (Salesforce): Tableau is the incumbent visual analytics platform with deep enterprise penetration. As Tableau adds AI-powered natural-language and agentic features (Tableau GPT/Einstein), it increasingly overlaps with Mantis's visual-first, AI-augmented data exploration thesis.
Market position
Strengths4 records
Weaknesses4 records
Competitive moat4 records
Key risks5 records
Key highlights6 records
Customer concentration
Mantis social profiles
Digital presenceMantis financial estimates
Financial estimateRevenue estimate
Valuation estimate
Mantis leadership team
Management profileNumber of profiles
Mantis funding detail
Funding detailFunding overview
Funding rounds
Investors
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Mantis M&A and investment
M&A and investmentM&A
Investments1 record
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Mantis
What does Mantis do?
Mantis is an interactive cognitive cartography platform that transforms multimodal knowledge (text, PDFs, tables, images, audio, video, coordinates, molecules, and biological data) into shared visual maps where people and AI reason together while preserving evidence, provenance, and missing context. The platform combines Semantic Spaces for visual data mapping, Composer and MantisGPT agents grounded in map context, Visual Workflows for node-based reusable analysis, and Developer Tools (MCP, CLI, developer API) for extensibility.
Is Mantis a public or private company?
Mantis is a private company. It is classified as unknown and is currently operating.
When was Mantis founded?
Mantis was founded in 2025. It employs 11 to 50 people.
Where is Mantis based?
Mantis is headquartered in Cambridge, United States, in the North America region.
How does Mantis make money?
One revenue line is on record: platform Access and Services.
Who are Mantis's main competitors?
Emerging players on record are Pinecone and Marimo. Direct peers are Weights & Biases, Observable, Hex, Noteable and Deepnote. Broad incumbents are Palantir Technologies, DataRobot and Tableau (Salesforce).
Does Mantis have an API?
Yes. Mantis offers a Developer API that enables extensibility. The platform also supports Mantis MCP (Model Context Protocol) tools, CLI for Codex, Claude Code, and OpenCode, custom panels and extensions, interactive notebooks, and webhooks for integration purposes. Developer documentation is at docs.withmantis.com.
What industry is Mantis in?
Mantis's product category is AI Data Analytics Platform. Its primary akta.pro industry code is HDAEANAH, Enterprise AI Data & Knowledge Platforms (Vector Databases, Knowledge Graphs), with a secondary code of BPAEAHAG, Generative AI & LLM Solutions Services (RAG, Agents, Copilots). Its NAICS code is 5132 and its SIC code is 7372.