Developer docs
API playgroundTry for free, no card

Search company profiles

Metaphi

Full company profile

uuid0003fw9

Namestring
Metaphi
Websiteurl
metaphi.ai
Company typeenum
Private
Founded yearint
2022
Descriptiontext

Metaphi is an AI infrastructure company that builds reinforcement learning (RL) environments and benchmarks for frontier AI companies developing autonomous coding and generation agents. The company's core product, Simhub, provides interactive simulation environments that move agent evaluation beyond static, non-interactive benchmarks, drawing on principles from autonomous vehicle simulation. Built on this platform are specialized public benchmarks including COBOLBench (100 enterprise COBOL maintenance tasks), FigmaBench (production Figma-to-code conversion), VideoBench (183 chapters across 41 courses for video/animation/presentation generation), and the LegacySWE research initiative on legacy enterprise system modernization. These products are delivered through the evals.metaphi.ai evaluation platform, which hosts a long-horizon agent leaderboard and targets frontier AI labs seeking realistic RL training and evaluation environments for out-of-distribution enterprise domains.

The company operates a research-first go-to-market motion, publishing technical essays and benchmarks as top-of-funnel demand generation, with enterprise sales conducted via direct contact forms capturing domain interest (Enterprise, Coding, Physical AI). Revenue is described as subscription-recurring from custom RL training environment contracts with frontier AI companies; pricing is not publicly disclosed and is quote-based. Founded in 2022 and based in McKinney, Texas, Metaphi has 1-10 employees, is led by co-founder Abhishek Chandwani, and is currently hiring RL engineers and ex-founders to extend into Physical AI and other unexplored domains. The company has raised $1.25M in seed funding from Flybridge (July 2022) with no subsequent rounds disclosed in available data.

Short descriptiontext

Metaphi builds reinforcement learning environments and specialized benchmarks (COBOLBench, FigmaBench, VideoBench) for frontier AI companies to train and evaluate autonomous coding and generation agents on enterprise-grade tasks.

Operating statusenum
Operating
Ownership categoryenum
Headcount rangeband
1–10
akta.pro rankint
HeadquartersMckinney, United States
HQ citystring
Mckinney
HQ countrystring
United States
HQ regionstring
North America
Markets served

Serves global market

Keyword5 values
RL training environments, AI agent evaluation, coding agent benchmarks, enterprise AI infrastructure, post-training platforms
Industry2 codes
PRIMARYHDAAALAJ
RLHF, Human Feedback & Evaluation Data Tools
Hardware, Technology & Telecom/Artificial Intelligence & Machine Learning/Synthetic Data, Data Labeling & Training Data Tools
SECONDARYHDAEANAA
End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management)
Hardware, Technology & Telecom/Enterprise Platforms & Databases/Artificial Intelligence Platforms for Enterprise
NAICS code1 code
  • Custom Computer Programming Services541511
SIC code1 code
  • Services-Prepackaged Software7372
Product category
AI Evaluation and RL Training Infrastructure
No data
GTM motion1 record

Each record includes

Type, Description, Source

Revenue model1 record
1Enterprise RL Environments
TypeSubscription Recurring
Description

Custom RL training environments built for frontier AI companies addressing out-of-distribution enterprise domains. Revenue derived from enterprise contracts for evaluation and training infrastructure.

metaphi.ai:443
Marketing channels3 records

Each record includes

Title, Type, Stage, Description, Source

Distribution channels1 record

Each record includes

Title, Type, Scope, Target buyer, Description, Source

Cost components3 values
Technology or R&D, Personnel, Marketing or Sales
GTM typeB2B
B2B
Offering typeSoftware
Software
Brand1 of 5 records shown
1Simhub
Description

Interactive simulation environments that move agent evaluation beyond static, non-interactive benchmarks.

metaphi.ai:443
+4 more records
Core offering1 text field

Metaphi builds reinforcement learning (RL) environments and evaluation benchmarks for frontier AI companies developing coding agents. The core offering is delivered through the evals.metaphi.ai platform, which provides interactive simulation environments (Simhub) and domain-specific benchmarks including COBOLBench (enterprise COBOL maintenance), FigmaBench (Figma-to-code conversion), and VideoBench (video/animation generation). These tools enable autonomous improvement of AI agents on out-of-distribution enterprise tasks beyond static, non-interactive benchmarks.

Differentiator
Functional benefit
Problem solved
Product overview1 text field

Metaphi builds RL (reinforcement learning) environments for frontier AI companies, offering a platform architecture centered on Simhub for interactive simulation environments. The core product is delivered through evals.metaphi.ai with two tiers: Enterprise for out-of-distribution enterprise domain evaluation, and Long-Horizon for extended task-sequence benchmarking with leaderboards. Specialized benchmarks include COBOLBench (100-task COBOL maintenance benchmark), VideoBench (video/animation generation from data rooms), and FigmaBench (Figma-to-code conversion). LegacySWE is a research initiative focused on legacy enterprise system maintenance. The platform tests AI coding agents through API interactions, design hierarchy extraction, and iterative deployment across enterprise domains including Coding, Enterprise, and Physical AI.

Product and service5 records
1Simhub
CategoryRL simulation platform
Description

Interactive simulation environments that move AI agent evaluation beyond static, non-interactive benchmarks. Built on principles from autonomous vehicle simulation to transform code generation agent training and evaluation for frontier AI companies.

2COBOLBench
CategoryBenchmark
Description

A 100-task public benchmark drawn from real enterprise COBOL maintenance environments, measuring frontier coding systems on legacy enterprise COBOL maintenance and modernization tasks. For frontier AI companies developing coding agents.

3VideoBench
CategoryBenchmark
Description

A benchmark measuring AI agents on video, animation, and presentation generation from curated data rooms, covering 183 chapters across 41 courses. For frontier AI companies developing video/animation generation agents.

4FigmaBench
CategoryBenchmark
Description

A benchmark measuring AI agents on production Figma-to-code conversion through API interaction, design hierarchy extraction, and iterative deployment. For frontier AI companies developing design-to-code conversion agents.

5Metaphi Evaluations Platform (evals.metaphi.ai)
CategoryEvaluation platform
Description

Enterprise AI evaluation platform providing RL environments and long-horizon agent leaderboards for measuring AI coding agents on out-of-distribution enterprise domains. Combines the Enterprise evaluation tier and Long-Horizon leaderboard tier as a single integrated platform.

Scale indicator2 records

Each record includes

Type, Value, Description, Source

Recent move6 records

Each record includes

Date, Type, Title, Description, Source

Expansion highlight5 records

Each record includes

Type, Description

Peers10 records
TypeDirect peer
Description

Provides an LLM evaluation, observability, and prompt engineering platform for AI teams building production applications. Directly comparable to Metaphi's enterprise evaluation platform (evals.metaphi.ai) and leaderboard-based approach to benchmarking AI agent quality.

TypeDirect peer
Description

Offers evaluation, tracing, and monitoring tooling for LLM-powered applications and agents. Comparable to Metaphi in serving AI engineering teams that need structured evaluation infrastructure beyond static benchmarks.

TypeDirect peer
Description

Builds an evaluation and observability platform specifically for LLM and generative AI applications. Overlaps with Metaphi's focus on agent evaluation, especially for production-grade AI systems.

TypeDirect peer
Description

Provides an LLM evaluation and experimentation platform aimed at teams shipping AI products. Comparable as a vendor helping enterprises evaluate and iterate on generative AI agents.

TypeBroad incumbent
Description

ML/LLM observability and evaluation platform serving production AI teams. Broader than Metaphi (covers traditional ML observability too), but directly competes in LLM/agent evaluation workflows.

TypeBroad incumbent
Description

Major RLHF, data labeling, and evaluation vendor serving frontier AI labs. Much larger and broader than Metaphi, but overlapping in supplying evaluation data and benchmark infrastructure to frontier model developers.

TypeBroad incumbent
Description

Data labeling and RLHF provider for frontier AI companies. Compares to Metaphi as a supplier of training/evaluation data infrastructure, though Surge is broader across modalities.

TypeBroad incumbent
Description

Established ML and AI development platform with evaluation, experiment tracking, and model registry capabilities. A broader incumbent whose eval tooling overlaps with Metaphi's agent evaluation focus.

9SWE-bench
TypeEmerging player
Description

Open-source benchmark for evaluating AI systems on real-world software engineering tasks. Directly comparable to Metaphi's COBOLBench and LegacySWE efforts, though focused on Python repos rather than legacy enterprise code.

TypeEmerging player
Description

Open-source LLM evaluation framework used by AI engineers to benchmark and test generative AI applications. Comparable to Metaphi as a vendor targeting teams that need structured agent/LLM evaluation.

Market position
Strengths4 records

Each record includes

Headline, Details, Source

Weaknesses4 records

Each record includes

Headline, Details, Source

Competitive moat4 records

Each record includes

Type, Details

Key risks6 records

Each record includes

Headline, Details, Source

Key highlights6 records

Each record includes

Headline, Details, Source

Customer concentration

Classification, Details

Segment2 records

Each record includes

Title, Type, Primary, Description, Pain point addressed, Use case, Source

Ideal customer profile1 record

Each record includes

Profile, Firmographic size, Sales motion, Sales cycle length, Buying structure, Purchase trigger, Buyer persona, Geography, Industry vertical, Primary use case, Description, Pain points, Evidence proof points, Target buyer

Technology focused
Yes
API detail
Has APIbool
No

Docs URL, Description

AI capability7 records

Each record includes

Type, Description, Source

AI maturity
App detail

Has app

Feature5 records

Each record includes

Title, Differentiator, Description, Source

Core technology
Revenue estimate
Valuation estimate
Number of profiles
Profiles1 record

Each record includes

Name, Designation, Designation category, Overview, Profile commentary, Source

No data
No data
Funding overview

Funding stage, Last funding date, Total funding USD

Funding rounds1 record

Each record includes

Round, Amount USD, Date, Pre money valuation, Total investors, Investors, News

Investors1 record

Each record includes

Name, Type, Date of entry, Rounds participated, Website

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Each record includes

Name, Acquisition type, Announced date, Completed date, Status, Website, News

Investment

Each record includes

Name, Round, Announced date, Lead investor, Website, News

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Metaphi

AI Evaluation and RL Training Infrastructuremetaphi.ai

Metaphi builds reinforcement learning environments and specialized benchmarks (COBOLBench, FigmaBench, VideoBench) for frontier AI companies to train and evaluate autonomous coding and generation agents on enterprise-grade tasks.

What Metaphi does

Metaphi is an AI infrastructure company that builds reinforcement learning (RL) environments and benchmarks for frontier AI companies developing autonomous coding and generation agents. The company's core product, Simhub, provides interactive simulation environments that move agent evaluation beyond static, non-interactive benchmarks, drawing on principles from autonomous vehicle simulation. Built on this platform are specialized public benchmarks including COBOLBench (100 enterprise COBOL maintenance tasks), FigmaBench (production Figma-to-code conversion), VideoBench (183 chapters across 41 courses for video/animation/presentation generation), and the LegacySWE research initiative on legacy enterprise system modernization. These products are delivered through the evals.metaphi.ai evaluation platform, which hosts a long-horizon agent leaderboard and targets frontier AI labs seeking realistic RL training and evaluation environments for out-of-distribution enterprise domains.

The company operates a research-first go-to-market motion, publishing technical essays and benchmarks as top-of-funnel demand generation, with enterprise sales conducted via direct contact forms capturing domain interest (Enterprise, Coding, Physical AI). Revenue is described as subscription-recurring from custom RL training environment contracts with frontier AI companies; pricing is not publicly disclosed and is quote-based. Founded in 2022 and based in McKinney, Texas, Metaphi has 1-10 employees, is led by co-founder Abhishek Chandwani, and is currently hiring RL engineers and ex-founders to extend into Physical AI and other unexplored domains. The company has raised $1.25M in seed funding from Flybridge (July 2022) with no subsequent rounds disclosed in available data.

Metaphi firmographics

Firmographics
Name
Metaphi
Website
https://metaphi.ai
Company type
Private
Founded year
2022
Operating status
Operating
Headcount range
1–10 employees
Short description
Metaphi builds reinforcement learning environments and specialized benchmarks (COBOLBench, FigmaBench, VideoBench) for frontier AI companies to train and evaluate autonomous coding and generation agents on enterprise-grade tasks.
Ownership category
akta.pro rank

Metaphi industry classification

Industry
Product category
AI Evaluation and RL Training Infrastructure
NAICS
Custom Computer Programming Services (541511)
SIC
Services-Prepackaged Software (7372)
akta.pro industry
PRIMARYHDAAALAJ
RLHF, Human Feedback & Evaluation Data Tools
Hardware, Technology & Telecom/Artificial Intelligence & Machine Learning/Synthetic Data, Data Labeling & Training Data Tools
SECONDARYHDAEANAA
End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management)
Hardware, Technology & Telecom/Enterprise Platforms & Databases/Artificial Intelligence Platforms for Enterprise

Keywords

  • RL training environments
  • AI agent evaluation
  • Coding agent benchmarks
  • Enterprise AI infrastructure
  • Post-training platforms

Where Metaphi is headquartered

Location

Headquarters

HQ city
Mckinney
HQ country
United States
HQ region
North America

Markets served

Metaphi business model

Business model
GTM type
B2B
Offering type
Software
Cost components
Technology or R&D, Personnel, Marketing or Sales

Revenue model

  1. Enterprise RL Environments: Custom RL training environments built for frontier AI companies addressing out-of-distribution enterprise domains. Revenue derived from enterprise contracts for evaluation and training infrastructure.

Go-to-market motion1 record

Distribution channels1 record

Marketing channels3 records

Metaphi product offering

Product offering

Core offering

Metaphi builds reinforcement learning (RL) environments and evaluation benchmarks for frontier AI companies developing coding agents. The core offering is delivered through the evals.metaphi.ai platform, which provides interactive simulation environments (Simhub) and domain-specific benchmarks including COBOLBench (enterprise COBOL maintenance), FigmaBench (Figma-to-code conversion), and VideoBench (video/animation generation). These tools enable autonomous improvement of AI agents on out-of-distribution enterprise tasks beyond static, non-interactive benchmarks.

Product overview

Metaphi builds RL (reinforcement learning) environments for frontier AI companies, offering a platform architecture centered on Simhub for interactive simulation environments. The core product is delivered through evals.metaphi.ai with two tiers: Enterprise for out-of-distribution enterprise domain evaluation, and Long-Horizon for extended task-sequence benchmarking with leaderboards. Specialized benchmarks include COBOLBench (100-task COBOL maintenance benchmark), VideoBench (video/animation generation from data rooms), and FigmaBench (Figma-to-code conversion). LegacySWE is a research initiative focused on legacy enterprise system maintenance. The platform tests AI coding agents through API interactions, design hierarchy extraction, and iterative deployment across enterprise domains including Coding, Enterprise, and Physical AI.

Differentiator

Problem solved

Functional benefit

Brands

  • Simhub: Interactive simulation environments that move agent evaluation beyond static, non-interactive benchmarks.
  • COBOLBench
  • FigmaBench
  • VideoBench
  • LegacySWE

Products and services

  • Simhub Interactive simulation environments that move AI agent evaluation beyond static, non-interactive benchmarks. Built on principles from autonomous vehicle simulation to transform code generation agent training and evaluation for frontier AI companies.
  • COBOLBench A 100-task public benchmark drawn from real enterprise COBOL maintenance environments, measuring frontier coding systems on legacy enterprise COBOL maintenance and modernization tasks. For frontier AI companies developing coding agents.
  • VideoBench A benchmark measuring AI agents on video, animation, and presentation generation from curated data rooms, covering 183 chapters across 41 courses. For frontier AI companies developing video/animation generation agents.
  • FigmaBench A benchmark measuring AI agents on production Figma-to-code conversion through API interaction, design hierarchy extraction, and iterative deployment. For frontier AI companies developing design-to-code conversion agents.
  • Metaphi Evaluations Platform (evals.metaphi.ai) Enterprise AI evaluation platform providing RL environments and long-horizon agent leaderboards for measuring AI coding agents on out-of-distribution enterprise domains. Combines the Enterprise evaluation tier and Long-Horizon leaderboard tier as a single integrated platform.

Companies that use Metaphi

Customer profile

Segments2 records

Ideal customer profiles1 record

Metaphi technology and API

Technology

Technology focussed Yes

API detail

Has API
No
API docs
API detail

Core technology

AI maturity

App detail

AI capability7 records

Feature5 records

Metaphi partnerships and signals

Strategic signal

Scale indicators2 records

Recent moves6 records

Expansion highlights5 records

Metaphi competitors and assessment

Company assessment

Direct peers

  • Braintrust: Provides an LLM evaluation, observability, and prompt engineering platform for AI teams building production applications. Directly comparable to Metaphi's enterprise evaluation platform (evals.metaphi.ai) and leaderboard-based approach to benchmarking AI agent quality.
  • LangSmith (LangChain): Offers evaluation, tracing, and monitoring tooling for LLM-powered applications and agents. Comparable to Metaphi in serving AI engineering teams that need structured evaluation infrastructure beyond static benchmarks.
  • Galileo: Builds an evaluation and observability platform specifically for LLM and generative AI applications. Overlaps with Metaphi's focus on agent evaluation, especially for production-grade AI systems.
  • HumanLoop: Provides an LLM evaluation and experimentation platform aimed at teams shipping AI products. Comparable as a vendor helping enterprises evaluate and iterate on generative AI agents.

Broad incumbents

  • Arize AI: ML/LLM observability and evaluation platform serving production AI teams. Broader than Metaphi (covers traditional ML observability too), but directly competes in LLM/agent evaluation workflows.
  • Scale AI: Major RLHF, data labeling, and evaluation vendor serving frontier AI labs. Much larger and broader than Metaphi, but overlapping in supplying evaluation data and benchmark infrastructure to frontier model developers.
  • Surge AI: Data labeling and RLHF provider for frontier AI companies. Compares to Metaphi as a supplier of training/evaluation data infrastructure, though Surge is broader across modalities.
  • Weights & Biases: Established ML and AI development platform with evaluation, experiment tracking, and model registry capabilities. A broader incumbent whose eval tooling overlaps with Metaphi's agent evaluation focus.

Emerging players

  • SWE-bench: Open-source benchmark for evaluating AI systems on real-world software engineering tasks. Directly comparable to Metaphi's COBOLBench and LegacySWE efforts, though focused on Python repos rather than legacy enterprise code.
  • Confident AI (DeepEval): Open-source LLM evaluation framework used by AI engineers to benchmark and test generative AI applications. Comparable to Metaphi as a vendor targeting teams that need structured agent/LLM evaluation.

Market position

Strengths4 records

Weaknesses4 records

Competitive moat4 records

Key risks6 records

Key highlights6 records

Customer concentration

Metaphi financial estimates

Financial estimate

Revenue estimate

Valuation estimate

Metaphi leadership team

Management profile

Number of profiles

Profiles1 record

Metaphi funding detail

Funding detail

Funding overview

Funding rounds1 record

Investors1 record

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

Metaphi M&A and investment

M&A and investment

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Metaphi

What does Metaphi do?

Metaphi builds reinforcement learning (RL) environments and evaluation benchmarks for frontier AI companies developing coding agents. The core offering is delivered through the evals.metaphi.ai platform, which provides interactive simulation environments (Simhub) and domain-specific benchmarks including COBOLBench (enterprise COBOL maintenance), FigmaBench (Figma-to-code conversion), and VideoBench (video/animation generation). These tools enable autonomous improvement of AI agents on out-of-distribution enterprise tasks beyond static, non-interactive benchmarks.

When was Metaphi founded?

Metaphi was founded in 2022. It employs 1 to 10 people.

Where is Metaphi based?

Metaphi is headquartered in Mckinney, United States, in the North America region.

How does Metaphi make money?

One revenue line is on record: enterprise RL Environments.

Who are Metaphi's main competitors?

Direct peers on record are Braintrust, LangSmith (LangChain), Galileo and HumanLoop. Broad incumbents are Arize AI, Scale AI, Surge AI and Weights & Biases. Emerging players are SWE-bench and Confident AI (DeepEval).

Does Metaphi have an API?

No public API is recorded for Metaphi.

What industry is Metaphi in?

Metaphi's product category is AI Evaluation and RL Training Infrastructure. Its primary akta.pro industry code is HDAAALAJ, RLHF, Human Feedback & Evaluation Data Tools, with a secondary code of HDAEANAA, End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management). Its NAICS code is 541511 and its SIC code is 7372.

Unlock the full company data

50 free credits on sign-up, no credit card required.

Contact sales