Developer docs
API playgroundTry for free, no card

Search company profiles

Moondream

Full company profile

uuid0000qsz

Namestring
Moondream
Legal namestring
M87 Labs, Inc.
Websiteurl
moondream.ai
Company typeenum
Private
Founded yearint
2024
Descriptiontext

Moondream, operating under legal entity M87 Labs, Inc., builds a family of small, production-grade vision language models and a complete deployment stack around them. The model line includes Moondream 3 Preview (a 9B-parameter sparse mixture-of-experts model with 2B active parameters and a 32K context window), Moondream 2 (a 2B dense workhorse with 5M+ monthly Hugging Face downloads), and Moondream 2 0.5B (a 500M-parameter distillation target for edge deployment). Native grounded skills (Detect, Point, Segment, Caption, Query) output structured spatial data — bounding boxes, coordinates, and SVG masks — rather than text descriptions.

The company surrounds its models with a vertically integrated platform: Lens for hosted fine-tuning via SFT and RL with as few as 20 labeled images, Photon as a proprietary inference engine reported ~2x faster than vLLM (34ms end-to-end on B200, with support for Apple Silicon, Windows, Jetson, and NVIDIA hardware from L4 to B200), Moondream Cloud as a HIPAA-compliant OpenAI-compatible hosted API at $0.30/$2.50 per 1M input/output tokens, and Moondream Station as a one-click desktop installer. Open weights are distributed freely on Hugging Face under a commercial-friendly license.

Monetization combines freemium open models with usage-based Cloud and Lens inference, a $350/month Team plan (10 seats, shared fine-tunes, dedicated Slack), and custom Enterprise contracts with HIPAA/BAA, SOC 2 in progress, and priority support. GTM is product-led and developer-centric via GitHub, Hugging Face, Discord, and documentation, supplemented by direct enterprise outreach. Named customers include PTZOptics (broadcast cameras) and CalPoly (academic research), with broader go-to-market targeted at developers, manufacturing and industrial, media and broadcasting, healthcare, retail, and security segments. Founded in 2024 by Jay (CEO) and Vik (CTO), both former AWS engineers with 8+ years of collaboration, the company has raised $4.5M in seed funding from Felicis Ventures, M12 (GitHub-Microsoft), and Ascend.

Short descriptiontext

Moondream builds small, production-grade open vision language models and a vertically integrated platform (Lens fine-tuning, Photon inference engine, Moondream Cloud API, Station desktop client) that targets developers, manufacturing, media/broadcast, healthcare, retail, and security teams needing fast, low-cost, real-time vision AI from edge to cloud.

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

Serves global market

Offices1 record

Each record includes

City, Country, Type, Description, Source

Keyword5 values
vision language models, open source AI, computer vision APIs, edge AI inference, model fine-tuning
Industry2 codes
PRIMARYHDAAACAA
Foundation Model Developers (LLM/Multimodal Model Labs)
Hardware, Technology & Telecom/Artificial Intelligence & Machine Learning/Foundation Models & Generative AI (LLMs, multimodal)
SECONDARYHDAAAJAB
On-Device Inference Runtimes & SDKs (mobile/embedded)
NAICS code2 codes
  • Software Publishers51321
  • Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services518210
SIC code1 code
  • Services-Prepackaged Software7372
Product category
Vision Language Models
Social media profiles2 records
GTM motion2 records

Each record includes

Type, Description, Source

Revenue model6 records
1Moondream Cloud (Hosted Inference)
TypeUsage Based
Description

Pay-per-token hosted inference. Real-time: $0.30/1M input tokens, $2.50/1M output tokens. Batch: $0.15/1M input, $1.25/1M output. Supports auto-scaling, HIPAA compliance, and OpenAI-compatible API.

moondream.ai
2Lens Fine-tuning
TypeUsage Based
Description

Token-based fine-tuning service. Training: $0.60/1M training tokens. Inference pricing same as Cloud.

moondream.ai
3Team Plan
TypeSubscription Recurring
Description

$350/month (or $300/mo billed annually). Includes 10 seats, shared fine-tunes, shared projects, shared API keys, shared billing, dedicated Slack, and 1h/month consulting.

moondream.ai
4Enterprise Plan
TypeSubscription Recurring
Description

Custom pricing with custom seats, HIPAA/BAA available, security review, priority support, custom deployment support, and custom billing.

moondream.ai
5Photon Inference Engine
TypeFreemium
Description

Free for local inference as of v1.3.0 (June 2026). Previously a paid product. Now completely free for self-hosted deployment.

moondream.ai
6Open Models
TypeFreemium
Description

Free open weights available on Hugging Face. Commercial-friendly licensing. No revenue directly generated but drives adoption and Cloud upsell.

moondream.ai
Marketing channels7 records

Each record includes

Title, Type, Stage, Description, Source

Distribution channels5 records

Each record includes

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

Cost components5 values
Technology or R&D, Personnel, Infrastructure, Marketing or Sales, Operations
Pricing details5 tiers
1Free tier for individual developers
ModelFreemiumBilling cadenceMonthly
Notes

$0/month. 1 seat, personal workspace, Discord support, $5/month usage credits. Photon and models are free.

moondream.ai
2Team collaboration and support
ModelSubscriptionBilling cadenceAnnual
Notes

$350/month or $300/mo billed annually. 10 seats, shared fine-tunes, shared projects, shared API keys, shared billing, dedicated Slack, 1h/month consulting. Additional seats available.

moondream.ai
3Enterprise custom deployment
ModelSubscriptionBilling cadenceMulti-year contract
Notes

Custom pricing. Custom seats, HIPAA/BAA available, security review, priority support, custom deployment support, custom billing.

moondream.ai
4Cloud pay-per-token inference
ModelUsage-basedBilling cadencePay-as-you-go
Notes

Real-time: $0.30/1M input tokens, $2.50/1M output tokens. Batch: $0.15/1M input, $1.25/1M output.

moondream.ai
5Lens fine-tuning training
ModelUsage-basedBilling cadencePay-as-you-go
Notes

$0.60/1M training tokens. Inference after training same as Cloud pricing.

moondream.ai
GTM typeB2B
B2B
Offering typeSoftware
Software
Core offering1 text field

Moondream (built by M87 Labs) develops and distributes a family of compact, open-weight vision language models (VLMs) including Moondream 3 Preview (9B MoE), Moondream 2 (2B dense), and Moondream 2 0.5B. The company sells a complete production stack around these models: the Photon high-performance inference engine, the Lens hosted fine-tuning service, Moondream Cloud hosted inference API, and the Moondream Station desktop client. Core revenue comes from pay-per-token cloud inference, token-based fine-tuning, and Team/Enterprise subscriptions with collaboration and support features.

Differentiator
Functional benefit
Problem solved
Quantifiable outcome1 of 8 values shown
  • F1 score improved from 0.28 to 0.79 for player with ball detection in NBA broadcasts (SFT fine-tuning)
+7 more records
Product overview1 text field

Moondream is a production vision language model platform built by M87 Labs, comprising a family of open-weight VLMs (Moondream 3 Preview, Moondream 2, Moondream 0.5B) and a complete production stack including Lens fine-tuning service, Photon inference engine, Moondream Cloud hosted API, and Moondream Station desktop client. The platform is designed for real-time vision AI with models that run everywhere from cloud servers to edge devices, offering built-in grounded skills for detection, captioning, query, pointing, and segmentation.

Product and service2 records
1Moondream 3 Preview
CategoryVision Language Model
2Moondream 2
Scale indicator10 records

Each record includes

Type, Value, Description, Source

Partnership1 partner
Strategic tierCoreTypeTechnology or IntegrationAnnounced on2026-02-19
Description

PTZOptics partnered with Moondream for visual reasoning initiative combining robotic PTZ camera systems with AI. Moondream's vision language model enables cameras to interpret what they see and trigger automated actions such as auto-tracking, smarter search, and real-time event-driven workflows. Supported by additional partners Axle AI, Detect-It, and LayerJot demonstrating visual AI applications.

Recent move7 records

Each record includes

Date, Type, Title, Description, Source

Expansion highlight7 records

Each record includes

Type, Description

Peers10 records
1LLaVA / LLaVA-Next
TypeDirect peer
Description

LLaVA is an open-source vision-language model from academic labs (originally Microsoft Research + UW) with permissive licensing. It is the closest direct peer to Moondream as an open-weight VLM competing on similar benchmarks and developer mindshare.

2Qwen-VL (Alibaba)
TypeDirect peer
Description

Qwen-VL is Alibaba's open-weight multimodal LLM family with strong vision-language benchmarks and permissive Apache licensing. Direct competitor to Moondream in the open-source VLM category targeting developers and production deployments.

3InternVL
TypeDirect peer
Description

InternVL is an open-source vision-language model family from Shanghai AI Lab that has achieved state-of-the-art results on multiple multimodal benchmarks. Direct peer competing in open-weight VLMs with comparable model sizes and developer adoption.

4Pixtral (Mistral AI)
TypeDirect peer
Description

Pixtral is Mistral AI's open-weight vision-language model with permissive Apache 2.0 licensing. Direct peer to Moondream as a production-grade open VLM from a well-funded AI lab targeting developers and enterprise customers.

TypeBroad incumbent
Description

OpenAI is the broadest incumbent in vision-language AI with GPT-4o and GPT-5 offering multimodal capabilities via API. Moondream's Cloud product directly competes for the same use cases, and OpenAI's API is the standard Moondream's API emulates.

TypeBroad incumbent
Description

Anthropic offers Claude with vision capabilities as part of its broader frontier LLM platform. Broad incumbent competing with Moondream for enterprise vision AI workloads where accuracy and reasoning quality matter more than cost.

TypeBroad incumbent
Description

Google's Gemini multimodal models are a broad incumbent offering vision-language capabilities as part of a wider AI platform, competing with Moondream for both developer adoption (via API) and enterprise vision AI workloads.

TypeEmerging player
Description

Landing AI is an Andrew Ng-led computer vision platform focused on industrial manufacturing and quality control use cases. Comparable to Moondream's manufacturing segment and fine-tuning approach, though specializing more narrowly in industrial inspection.

TypeEmerging player
Description

Roboflow provides computer vision tooling including model training, inference, and deployment infrastructure. Adjacent to Moondream's Lens fine-tuning and Photon inference engine offerings, targeting similar developer and enterprise vision AI workflows.

TypeEmerging player
Description

Ultralytics builds YOLO models and the Ultralytics HUB platform for training, deploying, and serving computer vision models. Comparable to Moondream in providing production vision model infrastructure with a developer-focused open-source ecosystem.

Market position
Strengths5 records

Each record includes

Headline, Details, Source

Weaknesses5 records

Each record includes

Headline, Details, Source

Competitive moat6 records

Each record includes

Type, Details

Key risks7 records

Each record includes

Headline, Details, Source

Key highlights7 records

Each record includes

Headline, Details, Source

Customer concentration

Classification, Details

Named customers2 records

Each record includes

Name, Industry, Type, Use case, Source, UUID

Segment6 records

Each record includes

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

Ideal customer profile2 records

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
Yes

Docs URL, Description

AI capability7 records

Each record includes

Type, Description, Source

AI maturity
App detail

Has app

Feature8 records

Each record includes

Title, Differentiator, Description, Source

Core technology
Revenue estimate
Valuation estimate
Number of profiles
Profiles2 records

Each record includes

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

No data
Compliance2 records

Each record includes

Name, Class, Description

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

Investors3 records

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 →

Moondream

Vision Language Modelsmoondream.ai

Moondream builds small, production-grade open vision language models and a vertically integrated platform (Lens fine-tuning, Photon inference engine, Moondream Cloud API, Station desktop client) that targets developers, manufacturing, media/broadcast, healthcare, retail, and security teams needing fast, low-cost, real-time vision AI from edge to cloud.

What Moondream does

Moondream, operating under legal entity M87 Labs, Inc., builds a family of small, production-grade vision language models and a complete deployment stack around them. The model line includes Moondream 3 Preview (a 9B-parameter sparse mixture-of-experts model with 2B active parameters and a 32K context window), Moondream 2 (a 2B dense workhorse with 5M+ monthly Hugging Face downloads), and Moondream 2 0.5B (a 500M-parameter distillation target for edge deployment). Native grounded skills (Detect, Point, Segment, Caption, Query) output structured spatial data — bounding boxes, coordinates, and SVG masks — rather than text descriptions.

The company surrounds its models with a vertically integrated platform: Lens for hosted fine-tuning via SFT and RL with as few as 20 labeled images, Photon as a proprietary inference engine reported ~2x faster than vLLM (34ms end-to-end on B200, with support for Apple Silicon, Windows, Jetson, and NVIDIA hardware from L4 to B200), Moondream Cloud as a HIPAA-compliant OpenAI-compatible hosted API at $0.30/$2.50 per 1M input/output tokens, and Moondream Station as a one-click desktop installer. Open weights are distributed freely on Hugging Face under a commercial-friendly license.

Monetization combines freemium open models with usage-based Cloud and Lens inference, a $350/month Team plan (10 seats, shared fine-tunes, dedicated Slack), and custom Enterprise contracts with HIPAA/BAA, SOC 2 in progress, and priority support. GTM is product-led and developer-centric via GitHub, Hugging Face, Discord, and documentation, supplemented by direct enterprise outreach. Named customers include PTZOptics (broadcast cameras) and CalPoly (academic research), with broader go-to-market targeted at developers, manufacturing and industrial, media and broadcasting, healthcare, retail, and security segments. Founded in 2024 by Jay (CEO) and Vik (CTO), both former AWS engineers with 8+ years of collaboration, the company has raised $4.5M in seed funding from Felicis Ventures, M12 (GitHub-Microsoft), and Ascend.

Moondream firmographics

Firmographics
Name
Moondream
Legal name
M87 Labs, Inc.
Website
https://moondream.ai
Company type
Private
Founded year
2024
Operating status
Operating
Headcount range
1–10 employees
Short description
Moondream builds small, production-grade open vision language models and a vertically integrated platform (Lens fine-tuning, Photon inference engine, Moondream Cloud API, Station desktop client) that targets developers, manufacturing, media/broadcast, healthcare, retail, and security teams needing fast, low-cost, real-time vision AI from edge to cloud.
Ownership category
akta.pro rank

Moondream industry classification

Industry
Product category
Vision Language Models
NAICS
Software Publishers (51321), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518210)
SIC
Services-Prepackaged Software (7372)
akta.pro industry
PRIMARYHDAAACAA
Foundation Model Developers (LLM/Multimodal Model Labs)
Hardware, Technology & Telecom/Artificial Intelligence & Machine Learning/Foundation Models & Generative AI (LLMs, multimodal)
SECONDARYHDAAAJAB
On-Device Inference Runtimes & SDKs (mobile/embedded)

Keywords

  • Vision language models
  • Open source AI
  • Computer vision APIs
  • Edge AI inference
  • Model fine-tuning

Where Moondream is headquartered

Location

Headquarters

HQ city
Seattle
HQ country
United States
HQ region
North America

Offices1 record

Markets served

Moondream business model

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

Revenue model

  1. Moondream Cloud (Hosted Inference): Pay-per-token hosted inference. Real-time: $0.30/1M input tokens, $2.50/1M output tokens. Batch: $0.15/1M input, $1.25/1M output. Supports auto-scaling, HIPAA compliance, and OpenAI-compatible API.
  2. Lens Fine-tuning: Token-based fine-tuning service. Training: $0.60/1M training tokens. Inference pricing same as Cloud.
  3. Team Plan: $350/month (or $300/mo billed annually). Includes 10 seats, shared fine-tunes, shared projects, shared API keys, shared billing, dedicated Slack, and 1h/month consulting.
  4. Enterprise Plan: Custom pricing with custom seats, HIPAA/BAA available, security review, priority support, custom deployment support, and custom billing.
  5. Photon Inference Engine: Free for local inference as of v1.3.0 (June 2026). Previously a paid product. Now completely free for self-hosted deployment.
  6. Open Models: Free open weights available on Hugging Face. Commercial-friendly licensing. No revenue directly generated but drives adoption and Cloud upsell.

Pricing tiers

ModelBillingPrice
FreemiumMonthlyFree tier for individual developers
SubscriptionAnnualTeam collaboration and support
SubscriptionMulti-year contractEnterprise custom deployment
Usage-basedPay-as-you-goCloud pay-per-token inference
Usage-basedPay-as-you-goLens fine-tuning training

Go-to-market motion2 records

Distribution channels5 records

Marketing channels7 records

Moondream product offering

Product offering

Core offering

Moondream (built by M87 Labs) develops and distributes a family of compact, open-weight vision language models (VLMs) including Moondream 3 Preview (9B MoE), Moondream 2 (2B dense), and Moondream 2 0.5B. The company sells a complete production stack around these models: the Photon high-performance inference engine, the Lens hosted fine-tuning service, Moondream Cloud hosted inference API, and the Moondream Station desktop client. Core revenue comes from pay-per-token cloud inference, token-based fine-tuning, and Team/Enterprise subscriptions with collaboration and support features.

Product overview

Moondream is a production vision language model platform built by M87 Labs, comprising a family of open-weight VLMs (Moondream 3 Preview, Moondream 2, Moondream 0.5B) and a complete production stack including Lens fine-tuning service, Photon inference engine, Moondream Cloud hosted API, and Moondream Station desktop client. The platform is designed for real-time vision AI with models that run everywhere from cloud servers to edge devices, offering built-in grounded skills for detection, captioning, query, pointing, and segmentation.

Differentiator

Problem solved

Functional benefit

Products and services

  • Moondream 3 Preview
  • Moondream 2

Quantifiable outcome

  • F1 score improved from 0.28 to 0.79 for player with ball detection in NBA broadcasts (SFT fine-tuning)
  • +7 more outcomes

Companies that use Moondream

Customer profile

Named customers2 records

Segments6 records

Ideal customer profiles2 records

Moondream technology and API

Technology

Technology focussed Yes

API detail

Has API
Yes
API docs
API detail

Core technology

AI maturity

App detail

AI capability7 records

Feature8 records

Moondream partnerships and signals

Strategic signal

Partnerships

One partnership is on record.

  • PTZOpticscoreTechnology or Integration · 19 February 2026PTZOptics partnered with Moondream for visual reasoning initiative combining robotic PTZ camera systems with AI. Moondream's vision language model enables cameras to interpret what they see and trigger automated actions such as auto-tracking, smarter search, and real-time event-driven workflows. Supported by additional partners Axle AI, Detect-It, and LayerJot demonstrating visual AI applications.

Scale indicators10 records

Recent moves7 records

Expansion highlights7 records

Moondream competitors and assessment

Company assessment

Direct peers

  • LLaVA / LLaVA-Next: LLaVA is an open-source vision-language model from academic labs (originally Microsoft Research + UW) with permissive licensing. It is the closest direct peer to Moondream as an open-weight VLM competing on similar benchmarks and developer mindshare.
  • Qwen-VL (Alibaba): Qwen-VL is Alibaba's open-weight multimodal LLM family with strong vision-language benchmarks and permissive Apache licensing. Direct competitor to Moondream in the open-source VLM category targeting developers and production deployments.
  • InternVL: InternVL is an open-source vision-language model family from Shanghai AI Lab that has achieved state-of-the-art results on multiple multimodal benchmarks. Direct peer competing in open-weight VLMs with comparable model sizes and developer adoption.
  • Pixtral (Mistral AI): Pixtral is Mistral AI's open-weight vision-language model with permissive Apache 2.0 licensing. Direct peer to Moondream as a production-grade open VLM from a well-funded AI lab targeting developers and enterprise customers.

Broad incumbents

  • OpenAI (GPT-4o / GPT-5): OpenAI is the broadest incumbent in vision-language AI with GPT-4o and GPT-5 offering multimodal capabilities via API. Moondream's Cloud product directly competes for the same use cases, and OpenAI's API is the standard Moondream's API emulates.
  • Anthropic (Claude with vision): Anthropic offers Claude with vision capabilities as part of its broader frontier LLM platform. Broad incumbent competing with Moondream for enterprise vision AI workloads where accuracy and reasoning quality matter more than cost.
  • Google (Gemini Vision): Google's Gemini multimodal models are a broad incumbent offering vision-language capabilities as part of a wider AI platform, competing with Moondream for both developer adoption (via API) and enterprise vision AI workloads.

Emerging players

  • Landing AI (LandingLens): Landing AI is an Andrew Ng-led computer vision platform focused on industrial manufacturing and quality control use cases. Comparable to Moondream's manufacturing segment and fine-tuning approach, though specializing more narrowly in industrial inspection.
  • Roboflow: Roboflow provides computer vision tooling including model training, inference, and deployment infrastructure. Adjacent to Moondream's Lens fine-tuning and Photon inference engine offerings, targeting similar developer and enterprise vision AI workflows.
  • Ultralytics: Ultralytics builds YOLO models and the Ultralytics HUB platform for training, deploying, and serving computer vision models. Comparable to Moondream in providing production vision model infrastructure with a developer-focused open-source ecosystem.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat6 records

Key risks7 records

Key highlights7 records

Customer concentration

Moondream social profiles

Digital presence

Moondream compliance and trust

Trust signal

Compliance2 records

Moondream financial estimates

Financial estimate

Revenue estimate

Valuation estimate

Moondream leadership team

Management profile

Number of profiles

Profiles2 records

Moondream funding detail

Funding detail

Funding overview

Funding rounds1 record

Investors3 records

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

Moondream 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 Moondream

What does Moondream do?

Moondream (built by M87 Labs) develops and distributes a family of compact, open-weight vision language models (VLMs) including Moondream 3 Preview (9B MoE), Moondream 2 (2B dense), and Moondream 2 0.5B. The company sells a complete production stack around these models: the Photon high-performance inference engine, the Lens hosted fine-tuning service, Moondream Cloud hosted inference API, and the Moondream Station desktop client. Core revenue comes from pay-per-token cloud inference, token-based fine-tuning, and Team/Enterprise subscriptions with collaboration and support features.

Is Moondream a public or private company?

Moondream is a private company. It is classified as venture growth investor backed and is currently operating.

When was Moondream founded?

Moondream was founded in 2024. It employs 1 to 10 people.

Where is Moondream based?

Moondream is headquartered in Seattle, United States, in the North America region.

How does Moondream make money?

Six revenue lines are on record. Moondream Cloud (Hosted Inference) is the primary driver. The others are lens Fine-tuning, team Plan, enterprise Plan, photon Inference Engine and open Models.

Who are Moondream's main competitors?

Direct peers on record are LLaVA / LLaVA-Next, Qwen-VL (Alibaba), InternVL and Pixtral (Mistral AI). Broad incumbents are OpenAI (GPT-4o / GPT-5), Anthropic (Claude with vision) and Google (Gemini Vision). Emerging players are Landing AI (LandingLens), Roboflow and Ultralytics.

Does Moondream have an API?

Yes. Moondream Cloud offers a hosted, OpenAI-compatible API for production vision language model inference. Developers can use the same SDK used for local inference, swapping in Cloud by setting runtime='cloud'. Supports captions, queries, detection, pointing, and segmentation. Free tier includes 2 requests/second; paid credits tier increases to 10 requests/second. Developer documentation is at docs.moondream.ai.

What industry is Moondream in?

Moondream's product category is Vision Language Models. Its primary akta.pro industry code is HDAAACAA, Foundation Model Developers (LLM/Multimodal Model Labs), with a secondary code of HDAAAJAB, On-Device Inference Runtimes & SDKs (mobile/embedded). Its NAICS code is 51321 and its SIC code is 7372.

Unlock the full company data

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

Contact sales
Live signals
RemioPrime Intellect Moondream Acquisition Puts Inference, Not Model Size, at the CenterPrime Intellect acquired Moondream, a vision-model startup, on October 13, 2026, to integrate its Photon inference engine and team. The deal shifts focus to inference, with Moondream Cloud shutting down on November 13, 2026, and customers migrating to Prime Inference. The acquisition tests whether integrated specialization beats general-purpose serving systems.GeekWirePrime Intellect acquires Moondream, a vision AI startup founded in Seattle by AWS vetsPrime Intellect acquired Seattle-based Moondream, a vision AI startup founded by AWS veterans, with terms undisclosed. Moondream has fewer than 10 employees and its cloud service will shut down Nov. 13, with customers moving to Prime Intellect. The acquisition aims to leverage Moondream's inference engine for faster model execution.Crypto BriefingPrime Intellect acquires Moondream to speed up AI inferencePrime Intellect acquired Moondream, an open-source multimodal AI company, to speed up AI inference at lower cost. The deal targets Moondream's GPU optimization software, including megakernel generation, with no financial terms or closing timeline disclosed. Prime Intellect's CEO cited robotics and computer-operating agents as key markets.PrimeintellectMoondream is joining Prime IntellectPrime Intellect acquired Moondream, with founders Vik Korrapati and Jay Allen joining to lead inference and multimodal model work. Moondream 3.1, a vision-language model, is now available through Prime Inference, powered by Photon, and Moondream Cloud customers will receive credits. The integration aims to combine training and fast inference for multimodal models.Sports Video GroupPTZOptics Launches Visual Reasoning Initiative, Partners with Moondream to Automate Video Decision-Making – Sports Video GroupPTZOptics announced the launch of its visual reasoning initiative, combining robotic PTZ camera systems with AI to make video actionable across broadcast, Pro AV, education, healthcare, and industrial environments. The company has partnered with Moondream, an open-source vision language model provider, to enable cameras to interpret what they see and trigger automated actions such as auto-tracking, smarter search, and real-time event-driven workflows. The initiative is supported by partner companies Axle AI, Detect-It, and LayerJot, which are demonstrating visual AI applications in media asset management, factory floor quality control, and surgical environments respectively.GeekWireWhy these startup founders are leaving Seattle for San FranciscoSeveral early-stage startup founders have relocated from Seattle to San Francisco during the AI boom, citing access to the 'world's AI capital,' denser startup networks, investor expectations, and a culture where startup work is normalized. Founders of companies including Talunt, MathGPT, Neuramill, and Moondream describe the Bay Area as offering better odds for success through serendipitous connections and funding opportunities. However, some Seattle founders and investors counter that the city remains a viable place to build, offering talent, lower costs, and a focus on sustainable growth.MoondreamTop Left: 2025-04-14 ReleaseMoondream released version 2025-04-14, a vision-language model trained on 450B tokens, improving document understanding and chart reasoning. Benchmarks improved on ChartQA (74.8 to 77.5) and DocVQA (76.5 to 79.3). Future releases will address repetition handling and tokenizer upgrades.MoondreamMoondream 2025-03-27 ReleaseMoondream released a new model on March 28, 2025, adding longer captions, improved object detection, OCR, and counting, and faster inference. The release also includes a planned mobile client update. The model is available for download and use in the playground.