MLflow
MLflow is an open-source AI engineering platform under the Linux Foundation that provides end-to-end tooling for developing, evaluating, deploying, and observing machine learning and LLM-based applications, used by thousands of organizations worldwide.
- Company typePrivate
- Founded2018
- HeadquartersSan Francisco, United States
- Headcount51–100
- GTM typeB2B
- OfferingSoftware
What MLflow does
MLflow is an open-source AI engineering platform initially created at Databricks in 2018 by Matei Zaharia and now governed by the Linux Foundation (Series of LF Projects, LLC) under the Apache 2.0 license. The platform provides end-to-end tooling for developing, evaluating, deploying, and observing machine learning and LLM-based applications. Its core products include experiment tracking, a model registry, an OpenTelemetry-native observability layer for traces and spans, a prompt registry with versioning and optimization, an AI Gateway for multi-provider LLM routing, evaluation tooling with 50+ metrics and LLM-as-judge scoring, multimodal and multi-turn tracing, MLflow Workspaces for collaboration, RBAC for governance, and a FastAPI-based Agent Server for deployment.
The platform is consumed by AI/ML engineering teams across thousands of organizations worldwide, with named enterprise adopters including Databricks, Microsoft, Meta, Toyota, Booking.com, Zillow, Wix, Accenture, and ASML. MLflow integrates with more than 100 AI frameworks and platforms — including LangChain, OpenAI, PyTorch, AWS SageMaker, Azure ML, Google Vertex AI, Anthropic, and TruLens — and is supported by a community of 900+ contributors, generating 30M+ monthly downloads and 20K+ GitHub stars. The Linux Foundation has governed the project for 5+ years, with Databricks continuing as the primary maintainer.
MLflow's business model is open-source distribution at zero direct cost: there is no pricing, no licensing fee, and no project-level revenue. Commercial value created by MLflow accrues primarily to Databricks through managed MLcloud offerings and to downstream vendors, rather than to the project entity itself. Growth is driven by community adoption, framework integration breadth, and feature velocity — exemplified by a dense 2026 product cadence covering the AI Gateway, Workspaces, multimodal tracing, multi-turn evaluation, and RBAC.
MLflow firmographics
Firmographics- Name
- MLflow
- Legal name
- MLflow Project, a Series of LF Projects, LLC
- Website
- https://mlflow.org
- Company type
- Private
- Founded year
- 2018
- Operating status
- Operating
- Headcount range
- 51–100 employees
- Short description
- MLflow is an open-source AI engineering platform under the Linux Foundation that provides end-to-end tooling for developing, evaluating, deploying, and observing machine learning and LLM-based applications, used by thousands of organizations worldwide.
- Ownership category
- akta.pro rank
MLflow industry classification
Industry- Product category
- MLOps/LLMOps Platform
- NAICS
- Software Publishers (5132)
- SIC
- Services-Prepackaged Software (7372)
- akta.pro primary industry
- End-to-End MLOps & ML Platform Suites (HDAAABAA)
- akta.pro secondary industries
- Pipeline Orchestration & Workflow Automation (HDAAABAD), Model Testing, Validation & Quality Assurance (HDAAABAH), MLOps/LLMOps & Model Lifecycle Management Services (BPAEAHAH)
Keywords
Where MLflow is headquartered
LocationHeadquarters
- HQ city
- San Francisco
- HQ country
- United States
- HQ region
- North America
Markets served
MLflow business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Technology or R&D, Infrastructure, Personnel, Operations
Revenue model
- Open Source Platform: MLflow is 100% open source under Apache 2.0 license, free for all users with no commercial licensing fees. Revenue is derived from parent company Databricks which provides enterprise-grade managed MLflow services.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Freemium | Monthly | Free Open Source - No paid tiers |
Go-to-market motion2 records
Distribution channels6 records
Marketing channels10 records
MLflow product offering
Product offeringCore offering
MLflow is an open-source AI engineering platform that supports the end-to-end machine learning and generative AI lifecycle. It provides integrated capabilities for LLM and agent observability (OpenTelemetry-native tracing), systematic evaluation with 50+ metrics and LLM judges, versioned prompt management with automated optimization, a unified AI Gateway across LLM providers, a FastAPI-based Agent Server, and traditional ML components including experiment tracking, model evaluation, standardized model packaging, and a Model Registry.
Product overview
MLflow is the largest open source AI engineering platform for agents, LLMs, and traditional ML models. The platform provides an integrated, full-lifecycle solution across two major areas: LLMs & Agents (observability, evaluation, prompt management, AI Gateway, and Agent Server) and Model Training (experiment tracking, model evaluation, MLflow Models packaging, and Model Registry). Built on Apache 2.0 license and governed by the Linux Foundation, MLflow enables teams to debug, evaluate, monitor, and optimize AI applications with 30+ million monthly downloads supporting Python, TypeScript/JavaScript, Java, and R.
Differentiator
Problem solved
Functional benefit
Products and services
- MLflow Observability Production-grade tracing built on OpenTelemetry that captures complete traces of LLM applications and agents, providing deep insights into behavior with latency, token counts, and cost monitoring for AI/ML engineers and enterprise AI teams.
- MLflow Evaluation Systematic evaluation framework with 50+ built-in metrics and LLM judges for running quality assessments, tracking metrics over time, and catching regressions before production, targeted at AI engineers and enterprise AI teams.
- Prompt Registry Versioned prompt management with full lineage tracking for testing and deploying prompts across environments, designed for prompt engineers and AI application teams.
- AI Gateway Unified API gateway providing OpenAI-compatible access to all LLM providers with rate limiting, cost controls, automatic fallback, and guardrails for content policies, intended for enterprise AI teams.
- Agent Server FastAPI-based hosting solution for deploying agents to production with automatic request validation, streaming support, and built-in tracing, designed for AI engineering teams.
- Experiment Tracking Core ML lifecycle management for tracking experiments, hyperparameters, and metrics for traditional machine learning and deep learning workflows, targeted at data scientists and ML practitioners.
- Model Evaluation Model evaluation capabilities for assessing ML model performance on test datasets with comparative analysis, for ML practitioners.
- MLflow Models Standardized model packaging format for deploying models to various serving targets with a consistent API, used by ML engineers and platform teams.
- Model Registry Centralized model management with versioning, stage transitions (staging/production/archived), and deployment capabilities, designed for ML platform teams.
Quantifiable outcome
- MemAlign achieved 74-89% reduction in judge error across misaligned evaluation dimensions
- +3 more outcomes
Companies that use MLflow
Customer profileNamed customers10 records
Segments4 records
Ideal customer profiles2 records
MLflow technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration22 records
AI capability14 records
Feature15 records
MLflow partnerships and signals
Strategic signalPartnerships
Ten partnerships are on record, tiered major and core.
- GooglemajorMLflow integrated into Google ADK (Agent Development Kit) as an observability platform with OpenTelemetry-native telemetry support, enabling ADK agents to use MLflow for tracing.
- Linux FoundationcoreLinux Foundation provides governance and stewardship for MLflow as a Series of LF Projects. Ensures open-source commitment and community governance.
- DatabrickscoreDatabricks is the primary maintainer of MLflow with core engineers dedicated to development. MLflow is deeply integrated into the Databricks platform.
- LangChainmajorMLflow integrates seamlessly with LangChain out of the box, supporting automatic tracing and evaluation for LangChain-based agents.
- OpenAImajorMLflow provides automatic tracing and evaluation support for OpenAI models and APIs through native autolog integration.
- PyTorchmajorMLflow integrates seamlessly with PyTorch for experiment tracking and model management in traditional ML workflows.
- AWS SageMakermajorMLflow supports deployment options on AWS SageMaker and integrates with SageMaker for end-to-end MLOps workflows.
- Azure Machine LearningmajorMLflow integrates with Azure Machine Learning as part of its cloud platform support for enterprise MLOps.
- Google Vertex AImajorMLflow integrated with Google Vertex AI for ML platform support and deployment options.
- AnthropicmajorMLflow AI Gateway supports Claude and Anthropic models with native integrations for observability and governance.
Scale indicators7 records
Recent moves6 records
Expansion highlights4 records
MLflow competitors and assessment
Company assessmentDirect peers
- Weights & Biases: Weights & Biases is a leading MLOps platform providing experiment tracking, model registry, and increasingly LLM evaluation and observability — competing head-on with MLflow's full-lifecycle ML and GenAI tooling.
- LangSmith: LangSmith is LangChain's commercial observability and evaluation platform for LLM applications and agents, offering tracing, evaluation, prompt management, and deployment — directly overlapping MLflow's core LLM/agent feature surface.
- Comet ML: Comet offers experiment tracking, model registry, and LLM evaluation tooling, positioning it as a direct alternative to MLflow for both classical ML lifecycle and GenAI observability workflows.
- Arize AI: Arize AI provides LLM tracing, evaluation, and observability, plus the open-source Phoenix project — directly comparable to MLflow's evaluation, tracing, and AI Gateway capabilities for production AI systems.
Emerging players
- WhyLabs: WhyLabs offers AI observability, LLM evaluation, and guardrails — an emerging player in production AI monitoring that competes with MLflow's evaluation, tracing, and governance features.
- Helicone: Helicone is an LLM observability and gateway platform offering tracing, prompt management, and cost controls, directly competing with MLflow's AI Gateway and tracing features.
- Portkey: Portkey provides an LLM gateway with observability, guardrails, and cost controls, overlapping directly with MLflow AI Gateway, budget policies, and guardrails.
- Langfuse: Langfuse is an open-source LLM observability and evaluation platform with tracing, prompt management, and scoring — closely comparable to MLflow's GenAI stack and an OSS-native alternative.
Broad incumbents
- Datadog: Datadog is a broad cloud observability platform with APM and LLM monitoring capabilities that increasingly overlap with MLflow's OpenTelemetry-based tracing and operational dashboards.
- Honeycomb.io: Honeycomb is an observability platform built on OpenTelemetry that supports ingesting MLflow traces for distributed AI application monitoring, making it a complementary but increasingly competitive infrastructure player.
Market position
Strengths5 records
Weaknesses4 records
Competitive moat5 records
Key risks5 records
Key highlights7 records
Customer concentration
MLflow social profiles
Digital presenceMLflow compliance and trust
Trust signalCompliance1 record
MLflow financial estimates
Financial estimateRevenue estimate
Valuation estimate
MLflow leadership team
Management profileNumber of profiles
MLflow funding detail
Funding detailFunding overview
Funding rounds
Investors
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
MLflow M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about MLflow
What does MLflow do?
MLflow is an open-source AI engineering platform that supports the end-to-end machine learning and generative AI lifecycle. It provides integrated capabilities for LLM and agent observability (OpenTelemetry-native tracing), systematic evaluation with 50+ metrics and LLM judges, versioned prompt management with automated optimization, a unified AI Gateway across LLM providers, a FastAPI-based Agent Server, and traditional ML components including experiment tracking, model evaluation, standardized model packaging, and a Model Registry.
Is MLflow a public or private company?
MLflow is a private company. It is classified as nonprofit foundation owned and is currently operating.
When was MLflow founded?
MLflow was founded in 2018. It employs 51 to 100 people.
Where is MLflow based?
MLflow is headquartered in San Francisco, United States, in the North America region.
How does MLflow make money?
One revenue line is on record: open Source Platform.
Who are MLflow's main competitors?
Direct peers on record are Weights & Biases, LangSmith, Comet ML and Arize AI. Emerging players are WhyLabs, Helicone, Portkey and Langfuse. Broad incumbents are Datadog and Honeycomb.io.
Does MLflow have an API?
Yes. MLflow provides a REST API and Python SDK for tracking experiments, managing models, logging traces, and accessing evaluation functions. The AI Gateway exposes a single OpenAI-compatible endpoint for routing requests to multiple LLM providers including OpenAI, Anthropic, Google Gemini, Amazon Bedrock, Azure OpenAI, Mistral, Cohere, DeepSeek, Groq, TogetherAI, xAI, OpenRouter, Ollama, and Vertex AI. The platform also offers an MCP (Model Context Protocol) server for connecting coding agents and LLM frameworks to MLflow. Developer documentation is at mlflow.org/docs/latest.
What industry is MLflow in?
MLflow's product category is MLOps/LLMOps Platform. Its primary akta.pro industry code is HDAAABAA, End-to-End MLOps & ML Platform Suites, with a secondary code of HDAAABAD, Pipeline Orchestration & Workflow Automation. Its NAICS code is 5132 and its SIC code is 7372.