01 About Reka
Reka is a frontier AI research and product company formed through the merger of Reka AI and Moonvalley. The combined team brings together multimodal research and video intelligence expertise under one mission: building AI for the physical world. Reka's models are trained to perceive and reason across text, images, video, and audio at the same time, so machines can understand and act on the full complexity of real-world environments rather than text alone. Reka's tagline is Multimodal AI you can deploy anywhere.
Reka is a frontier AI research and product company formed through the merger of Reka AI and Moonvalley. The combined team brings together multimodal research and video intelligence expertise under one mission: building AI for the physical world. Reka's models are trained to perceive and reason across text, images, video, and audio at the same time, so machines can understand and act on the full complexity of real-world environments rather than text alone. Reka's tagline is Multimodal AI you can deploy anywhere.
Reka was founded by a team of AI researchers and engineers from Google DeepMind, Meta AI, Google Brain, and FAIR (Meta's Fundamental AI Research lab). The company is headquartered in San Francisco, California.
Reka was founded by a team of AI researchers and engineers from Google DeepMind, Meta AI, Google Brain, and FAIR (Meta's Fundamental AI Research lab). The company is headquartered in San Francisco, California.
Reka has raised $170 million in funding and is valued at $1.3 billion. Its investors and strategic partners include NVIDIA, Snowflake, Radical Ventures, and DST Global. Having NVIDIA and Snowflake on board as strategic partners reflects Reka's position at the meeting point of frontier AI research and enterprise deployment infrastructure.
Reka has raised $170 million in funding and is valued at $1.3 billion. Its investors and strategic partners include NVIDIA, Snowflake, Radical Ventures, and DST Global. Having NVIDIA and Snowflake on board as strategic partners reflects Reka's position at the meeting point of frontier AI research and enterprise deployment infrastructure.
Reka operates with two goals: push the frontier of physical world understanding, and translate that research into useful, enterprise-ready AI products. Reka's view is that multimodal capabilities, meaning the ability to reason across text, image, video, and audio at the same time, are foundational to the next generation of AI intelligence. Reka focuses on algorithmic breakthroughs rather than brute-force scaling, achieving competitive performance at a fraction of the cost of larger systems.
Reka operates with two goals: push the frontier of physical world understanding, and translate that research into useful, enterprise-ready AI products. Reka's view is that multimodal capabilities, meaning the ability to reason across text, image, video, and audio at the same time, are foundational to the next generation of AI intelligence. Reka focuses on algorithmic breakthroughs rather than brute-force scaling, achieving competitive performance at a fraction of the cost of larger systems.
02 Models & Technology
Four differentiators set Reka apart from other AI providers:
- Multimodal by design:Reka's models are trained from day one on video, image, audio, and text. They are not adapted from text-only systems. This is a fundamental architectural difference, not a feature layer.
- Intelligence for the physical world:Reka builds AI that reasons about dynamic, real-world environments such as live video streams, industrial sensors, and surveillance networks, not just documents and text.
- Algorithmic efficiency:Reka reaches frontier-level performance through architectural innovation rather than brute-force compute scaling, which makes deployment viable at enterprise cost structures.
- Deployment flexibility:Reka models can be deployed in SaaS, VPC, on-premise, and fully air-gapped environments. This meets the compliance requirements of defence, government, and regulated industries that cloud-only providers cannot serve.
Four differentiators set Reka apart from other AI providers:
- Multimodal by design:Reka's models are trained from day one on video, image, audio, and text. They are not adapted from text-only systems. This is a fundamental architectural difference, not a feature layer.
- Intelligence for the physical world:Reka builds AI that reasons about dynamic, real-world environments such as live video streams, industrial sensors, and surveillance networks, not just documents and text.
- Algorithmic efficiency:Reka reaches frontier-level performance through architectural innovation rather than brute-force compute scaling, which makes deployment viable at enterprise cost structures.
- Deployment flexibility:Reka models can be deployed in SaaS, VPC, on-premise, and fully air-gapped environments. This meets the compliance requirements of defence, government, and regulated industries that cloud-only providers cannot serve.
Reka offers a tiered family of natively multimodal foundation models, each designed for different performance, cost, and deployment requirements:
- Reka Core:The flagship frontier model. Built for high-stakes, complex multimodal reasoning tasks that require maximum accuracy across large, mixed-media inputs.
- Reka Flash:A 21 billion parameter model that balances speed, cost, and performance. The primary model for enterprise-grade production applications where throughput and efficiency are priorities.
- Reka Edge:A compact model ranging from 1 to 7 billion parameters, optimised for on-device and edge deployments including robotics, IoT, and mobile scenarios that require low latency and a small compute footprint.
- Reka Spark:A lightweight model in the Edge family, optimised for high-efficiency inference on constrained hardware.
Reka offers a tiered family of natively multimodal foundation models, each designed for different performance, cost, and deployment requirements:
- Reka Core:The flagship frontier model. Built for high-stakes, complex multimodal reasoning tasks that require maximum accuracy across large, mixed-media inputs.
- Reka Flash:A 21 billion parameter model that balances speed, cost, and performance. The primary model for enterprise-grade production applications where throughput and efficiency are priorities.
- Reka Edge:A compact model ranging from 1 to 7 billion parameters, optimised for on-device and edge deployments including robotics, IoT, and mobile scenarios that require low latency and a small compute footprint.
- Reka Spark:A lightweight model in the Edge family, optimised for high-efficiency inference on constrained hardware.
Reka Core is the frontier-class model for the most complex and high-stakes reasoning tasks. It processes massive, mixed-modality inputs with maximum accuracy. Reka Flash is a highly optimised 21 billion parameter model that prioritises throughput and cost-effectiveness, which makes it the practical choice for most production enterprise applications. Both models are natively multimodal.
Reka Core is the frontier-class model for the most complex and high-stakes reasoning tasks. It processes massive, mixed-modality inputs with maximum accuracy. Reka Flash is a highly optimised 21 billion parameter model that prioritises throughput and cost-effectiveness, which makes it the practical choice for most production enterprise applications. Both models are natively multimodal.
Reka Vision uses a multi-frame analysis approach that consolidates multiple video frames at the same time. This contrasts with conventional solutions that analyse frames independently, which results in lower accuracy and slower insights. By processing temporal context across frames in parallel, Reka Vision achieves lower latency and higher accuracy for complex event detection and natural language video search.
Reka Vision uses a multi-frame analysis approach that consolidates multiple video frames at the same time. This contrasts with conventional solutions that analyse frames independently, which results in lower accuracy and slower insights. By processing temporal context across frames in parallel, Reka Vision achieves lower latency and higher accuracy for complex event detection and natural language video search.
03 Products & Platforms
Yes. Reka Edge and Reka Spark are purpose-built for local hardware deployment, running on robotics platforms, mobile devices, and IoT sensors with minimal compute requirements. Reka Speech is similarly designed for on-device transcription and translation, providing real-time performance without cloud round-trips. All edge models are available both on-device and via API.
Yes. Reka Edge and Reka Spark are purpose-built for local hardware deployment, running on robotics platforms, mobile devices, and IoT sensors with minimal compute requirements. Reka Speech is similarly designed for on-device transcription and translation, providing real-time performance without cloud round-trips. All edge models are available both on-device and via API.
Reka's product suite is available at reka.ai and is organised into two tiers: production-ready platforms for enterprise deployment, and playground tools for development and exploration.
Production platforms:
- Vision Platform:Reka's primary enterprise offering. An enterprise video intelligence API for indexing, searching, and analysing any video at scale. Available at vision.reka.ai.
- Reka Clip:Built for creators. Turns long videos into viral clips, captions, and highlights with one click.
Playground tools (available for development and testing):
- Reka Edge:Test the latest multimodal models on image and video in real time.
- Reka Research:AI built for breakthroughs. Navigates large knowledge bases across public and private sources.
- Reka Chat:Fast, grounded, multimodal answers at enterprise scale.
- Reka Speech:Real-time speech AI for multilingual interaction.
Reka's product suite is available at reka.ai and is organised into two tiers: production-ready platforms for enterprise deployment, and playground tools for development and exploration.
Production platforms:
- Vision Platform:Reka's primary enterprise offering. An enterprise video intelligence API for indexing, searching, and analysing any video at scale. Available at vision.reka.ai.
- Reka Clip:Built for creators. Turns long videos into viral clips, captions, and highlights with one click.
Playground tools (available for development and testing):
- Reka Edge:Test the latest multimodal models on image and video in real time.
- Reka Research:AI built for breakthroughs. Navigates large knowledge bases across public and private sources.
- Reka Chat:Fast, grounded, multimodal answers at enterprise scale.
- Reka Speech:Real-time speech AI for multilingual interaction.
The Vision Platform is Reka's primary enterprise product. It is a video intelligence API that lets organisations index, search, and analyse any video at scale. Built on Reka's state-of-the-art multimodal foundation models, it supports natural language querying across video archives and real-time streams. The Vision Platform is production-ready and available for enterprise deployment at vision.reka.ai.
The Vision Platform is Reka's primary enterprise product. It is a video intelligence API that lets organisations index, search, and analyse any video at scale. Built on Reka's state-of-the-art multimodal foundation models, it supports natural language querying across video archives and real-time streams. The Vision Platform is production-ready and available for enterprise deployment at vision.reka.ai.
Reka Clip is a short-form content creation platform built for creators and media teams. It turns long-form video such as podcasts, livestreams, broadcasts, and sports coverage into viral clips, captions, and highlights with a single click. Reka Clip handles auto-reframing for vertical video formats (9:16) and generates accurate captions. It is production-ready and Reka's dedicated creator-focused platform.
Reka Clip is a short-form content creation platform built for creators and media teams. It turns long-form video such as podcasts, livestreams, broadcasts, and sports coverage into viral clips, captions, and highlights with a single click. Reka Clip handles auto-reframing for vertical video formats (9:16) and generates accurate captions. It is production-ready and Reka's dedicated creator-focused platform.
Reka Chat is a multimodal conversational AI product that delivers fast, grounded answers at enterprise scale. Unlike text-only chatbots, Reka Chat is built on Reka's natively multimodal architecture, which lets it reason across text, images, video, and audio inputs. It is available as a playground tool at reka.ai.
Reka Chat is a multimodal conversational AI product that delivers fast, grounded answers at enterprise scale. Unlike text-only chatbots, Reka Chat is built on Reka's natively multimodal architecture, which lets it reason across text, images, video, and audio inputs. It is available as a playground tool at reka.ai.
Reka Research is an agentic AI system built for breakthroughs. It is designed to navigate large knowledge bases across both public web sources and private documents, and it provides cited, traceable outputs for financial research, legal discovery, compliance monitoring, and enterprise knowledge management. Available as a playground tool at reka.ai.
Reka Research is an agentic AI system built for breakthroughs. It is designed to navigate large knowledge bases across both public web sources and private documents, and it provides cited, traceable outputs for financial research, legal discovery, compliance monitoring, and enterprise knowledge management. Available as a playground tool at reka.ai.
Reka Speech is a real-time speech AI platform for multilingual interaction. It offers ultra-compact, low-latency transcription and translation, deployable on-device or via API. Designed for high-throughput voice workflows, Reka Speech supports multilingual use cases across enterprise and edge environments.
Reka Speech is a real-time speech AI platform for multilingual interaction. It offers ultra-compact, low-latency transcription and translation, deployable on-device or via API. Designed for high-throughput voice workflows, Reka Speech supports multilingual use cases across enterprise and edge environments.
04 Use Cases & Industries
Reka Edge is Reka's compact model family (1 to 7 billion parameters) for on-device and edge deployment. The Edge playground lets developers test Reka's latest models on image and video in real time. Edge models are optimised for robotics, IoT, and mobile scenarios where low latency and a small compute footprint are critical.
Reka Edge is Reka's compact model family (1 to 7 billion parameters) for on-device and edge deployment. The Edge playground lets developers test Reka's latest models on image and video in real time. Edge models are optimised for robotics, IoT, and mobile scenarios where low latency and a small compute footprint are critical.
Reka's models are general-purpose but are particularly well-suited to industries that handle high volumes of complex, unstructured multimodal data:
- Physical Security & Public Safety:Real-time event detection, natural language search across camera networks, and automated incident reporting for law enforcement and enterprise security teams.
- Media, Sports Broadcasting & Entertainment:Automated highlight generation, contextual clip creation, sponsorship verification, ad slot optimisation, and archive monetisation at scale.
- Financial Services:Deep research, multi-source document synthesis, compliance monitoring, and investment intelligence.
- Enterprise Operations:Safety compliance monitoring, anomaly detection, perimeter security, and incident investigation across manufacturing, logistics, healthcare, and critical infrastructure.
- E-commerce & Retail:Visual search, product discovery, and consumer trend analysis from image and video data.
- Robotics & Industrial:Vision-based anomaly detection and quality control deployed directly at the edge.
- Government & Defence:Air-gapped deployment for sovereign AI requirements, public safety monitoring, and mission-critical video intelligence.
Reka's models are general-purpose but are particularly well-suited to industries that handle high volumes of complex, unstructured multimodal data:
- Physical Security & Public Safety:Real-time event detection, natural language search across camera networks, and automated incident reporting for law enforcement and enterprise security teams.
- Media, Sports Broadcasting & Entertainment:Automated highlight generation, contextual clip creation, sponsorship verification, ad slot optimisation, and archive monetisation at scale.
- Financial Services:Deep research, multi-source document synthesis, compliance monitoring, and investment intelligence.
- Enterprise Operations:Safety compliance monitoring, anomaly detection, perimeter security, and incident investigation across manufacturing, logistics, healthcare, and critical infrastructure.
- E-commerce & Retail:Visual search, product discovery, and consumer trend analysis from image and video data.
- Robotics & Industrial:Vision-based anomaly detection and quality control deployed directly at the edge.
- Government & Defence:Air-gapped deployment for sovereign AI requirements, public safety monitoring, and mission-critical video intelligence.
Physical AI refers to AI systems that can not only understand video and sensor data, but use that understanding to reason about and interact with the physical world in real time. Reka positions video intelligence as the foundational infrastructure layer for Physical AI. Early deployments include autonomous inspection systems that identify equipment anomalies without human review, warehouse environments where AI monitors safety compliance continuously across hundreds of zones at the same time, and operational workflows where visual understanding triggers downstream actions without manual intervention.
Physical AI refers to AI systems that can not only understand video and sensor data, but use that understanding to reason about and interact with the physical world in real time. Reka positions video intelligence as the foundational infrastructure layer for Physical AI. Early deployments include autonomous inspection systems that identify equipment anomalies without human review, warehouse environments where AI monitors safety compliance continuously across hundreds of zones at the same time, and operational workflows where visual understanding triggers downstream actions without manual intervention.
For media and sports broadcasting organisations, Reka Vision addresses the challenge of extracting commercial value from both live feeds and large historical archives:
- Intelligence at the speed of live:Reka Vision understands sport-specific gameplay visually, without relying on audio commentary. It recognises emotional peaks, tracks individual players by jersey number, and generates highlight packages ready for social platforms by the final whistle.
- Sponsorship and ad revenue verification:Frame-level, timestamped proof-of-delivery reports track brand logos across every camera angle, even during motion blur or partial occlusions, which lets broadcasters charge a premium for high-excitement exposure.
- Archive monetisation:Reka Vision turns stored footage from a cost centre into a searchable, licensable inventory. Every archived asset becomes findable and therefore commercially valuable.
- Ad slot optimisation:Reka Vision identifies optimal moments to place advertisements based on content context and emotional tone rather than fixed time intervals. This maximises engagement while keeping brand safety intact.
For media and sports broadcasting organisations, Reka Vision addresses the challenge of extracting commercial value from both live feeds and large historical archives:
- Intelligence at the speed of live:Reka Vision understands sport-specific gameplay visually, without relying on audio commentary. It recognises emotional peaks, tracks individual players by jersey number, and generates highlight packages ready for social platforms by the final whistle.
- Sponsorship and ad revenue verification:Frame-level, timestamped proof-of-delivery reports track brand logos across every camera angle, even during motion blur or partial occlusions, which lets broadcasters charge a premium for high-excitement exposure.
- Archive monetisation:Reka Vision turns stored footage from a cost centre into a searchable, licensable inventory. Every archived asset becomes findable and therefore commercially valuable.
- Ad slot optimisation:Reka Vision identifies optimal moments to place advertisements based on content context and emotional tone rather than fixed time intervals. This maximises engagement while keeping brand safety intact.
05 Customers & Partners
In manufacturing, logistics, healthcare, and critical infrastructure, Reka Vision closes the gap between what gets recorded and what gets understood. Operators can run natural language queries across entire camera networks in seconds, finding every instance of a worker entering a restricted zone without a safety vest, or identifying vehicles that stopped near a loading bay overnight. Reka Vision delivers a 95% reduction in false alarms for enterprise operations and compresses post-incident documentation from hours to minutes through automated report generation.
In manufacturing, logistics, healthcare, and critical infrastructure, Reka Vision closes the gap between what gets recorded and what gets understood. Operators can run natural language queries across entire camera networks in seconds, finding every instance of a worker entering a restricted zone without a safety vest, or identifying vehicles that stopped near a loading bay overnight. Reka Vision delivers a 95% reduction in false alarms for enterprise operations and compresses post-incident documentation from hours to minutes through automated report generation.
Reka is deployed by organisations across media, public safety, and enterprise operations globally. Notable customers include:
- Shutterstock:Uses Reka Vision to enrich metadata and tagging across its library of over 50 million videos, turning stored content into a searchable, licensable commercial asset.
- Turing Video:Built Guardian AI, an agentic surveillance platform, on top of Reka Vision. Deployed by law enforcement and enterprise security teams across the United States, running entirely within customer infrastructure with no footage leaving the secure environment.
- Orange Village Police Department, Ohio:Uses Reka Vision to streamline video investigation and public safety monitoring.
Reka is deployed by organisations across media, public safety, and enterprise operations globally. Notable customers include:
- Shutterstock:Uses Reka Vision to enrich metadata and tagging across its library of over 50 million videos, turning stored content into a searchable, licensable commercial asset.
- Turing Video:Built Guardian AI, an agentic surveillance platform, on top of Reka Vision. Deployed by law enforcement and enterprise security teams across the United States, running entirely within customer infrastructure with no footage leaving the secure environment.
- Orange Village Police Department, Ohio:Uses Reka Vision to streamline video investigation and public safety monitoring.
Reka works with a network of technology, infrastructure, and go-to-market partners that extend its reach across enterprise and government markets:
- NVIDIA:Strategic investor and technology partner. Reka's models are optimised for NVIDIA infrastructure, which reflects deep alignment on the future of accelerated, multimodal AI computing.
- HPE (Hewlett Packard Enterprise):Infrastructure partner enabling Reka's deployment across HPE's enterprise server and edge computing environments. Particularly relevant for organisations that need on-premise and air-gapped deployments.
- Deloitte:Consulting and go-to-market partner. Deloitte's global enterprise and government relationships extend Reka's reach into large-scale digital transformation programmes across regulated industries.
- Carahsoft:A leading US government IT solutions provider and distributor. The Carahsoft partnership lets Reka serve US federal, state, and local government agencies and defence organisations through established procurement channels.
- Oracle:Distribution partner. Reka models are available via Oracle Cloud Infrastructure (OCI), which makes them accessible to Oracle's global enterprise customer base through existing OCI procurement.
Reka works with a network of technology, infrastructure, and go-to-market partners that extend its reach across enterprise and government markets:
- NVIDIA:Strategic investor and technology partner. Reka's models are optimised for NVIDIA infrastructure, which reflects deep alignment on the future of accelerated, multimodal AI computing.
- HPE (Hewlett Packard Enterprise):Infrastructure partner enabling Reka's deployment across HPE's enterprise server and edge computing environments. Particularly relevant for organisations that need on-premise and air-gapped deployments.
- Deloitte:Consulting and go-to-market partner. Deloitte's global enterprise and government relationships extend Reka's reach into large-scale digital transformation programmes across regulated industries.
- Carahsoft:A leading US government IT solutions provider and distributor. The Carahsoft partnership lets Reka serve US federal, state, and local government agencies and defence organisations through established procurement channels.
- Oracle:Distribution partner. Reka models are available via Oracle Cloud Infrastructure (OCI), which makes them accessible to Oracle's global enterprise customer base through existing OCI procurement.
Organisations deploying Reka's AI platforms report significant and measurable operational improvements:
- 65% faster case resolution for public safety organisations using Reka Vision, while maintaining full air-gapped data sovereignty.
- 95% reduction in false alarms for enterprise operations teams using Reka Vision for perimeter monitoring and safety compliance.
- 42% reduction in crime rates in public safety deployments of Reka Vision.
- 40% reduction in operational costs related to manual video and data processing workflows.
- Shutterstock turned 50 million archived videos from unsearchable storage into a fully licensable, revenue-generating content inventory.
Organisations deploying Reka's AI platforms report significant and measurable operational improvements:
- 65% faster case resolution for public safety organisations using Reka Vision, while maintaining full air-gapped data sovereignty.
- 95% reduction in false alarms for enterprise operations teams using Reka Vision for perimeter monitoring and safety compliance.
- 42% reduction in crime rates in public safety deployments of Reka Vision.
- 40% reduction in operational costs related to manual video and data processing workflows.
- Shutterstock turned 50 million archived videos from unsearchable storage into a fully licensable, revenue-generating content inventory.
06 Deployment & Integration
Paul Hennessy, CEO of Shutterstock, said that Shutterstock is using Reka's AI capabilities to further enrich metadata and tagging of its content library. Christopher Zenaty, President of Turing Video, described Reka Vision as the best platform to build on, citing its visual understanding capabilities and ability to scale to tens of millions of daily events.
Paul Hennessy, CEO of Shutterstock, said that Shutterstock is using Reka's AI capabilities to further enrich metadata and tagging of its content library. Christopher Zenaty, President of Turing Video, described Reka Vision as the best platform to build on, citing its visual understanding capabilities and ability to scale to tens of millions of daily events.
Reka offers four deployment models to meet diverse infrastructure and compliance requirements:
- Reka Managed Cloud (SaaS):Fastest onboarding via Reka's secure API. No infrastructure required.
- Virtual Private Cloud (VPC):Deploy Reka models within your own AWS, GCP, or Azure environment. Data never leaves your security perimeter.
- On-Premise:Full local deployment on customer hardware within the organisation's own network.
- Air-Gapped:For defence, government, and critical infrastructure. All processing happens within a secure perimeter with zero external connectivity.
Reka offers four deployment models to meet diverse infrastructure and compliance requirements:
- Reka Managed Cloud (SaaS):Fastest onboarding via Reka's secure API. No infrastructure required.
- Virtual Private Cloud (VPC):Deploy Reka models within your own AWS, GCP, or Azure environment. Data never leaves your security perimeter.
- On-Premise:Full local deployment on customer hardware within the organisation's own network.
- Air-Gapped:For defence, government, and critical infrastructure. All processing happens within a secure perimeter with zero external connectivity.
Data sovereignty is a core design requirement for Reka, not a feature. For enterprises in regulated industries, where video data is processed is a regulatory requirement, not a preference. Reka's architecture supports every deployment model that compliance teams require, including fully air-gapped environments where no data leaves the organisation's infrastructure. This is the precondition for deployment in defence, healthcare, financial services, and government sectors.
Data sovereignty is a core design requirement for Reka, not a feature. For enterprises in regulated industries, where video data is processed is a regulatory requirement, not a preference. Reka's architecture supports every deployment model that compliance teams require, including fully air-gapped environments where no data leaves the organisation's infrastructure. This is the precondition for deployment in defence, healthcare, financial services, and government sectors.
Yes. Reka's models and platforms are designed for enterprise integration without requiring replacement of existing camera or storage infrastructure. Reka Vision scales linearly with camera count, which lets organisations grow alongside their existing systems. The intelligence layer sits on top of current infrastructure, so organisations do not need to replace existing investments to deploy Reka.
Yes. Reka's models and platforms are designed for enterprise integration without requiring replacement of existing camera or storage infrastructure. Reka Vision scales linearly with camera count, which lets organisations grow alongside their existing systems. The intelligence layer sits on top of current infrastructure, so organisations do not need to replace existing investments to deploy Reka.
There are two ways to get started. Organisations can upload footage directly to the Vision Platform at vision.reka.ai and run natural language queries against their own content with no sales conversation required. For organisations ready to scope a pilot or discuss deployment options for specific infrastructure requirements, a demo can be requested at reka.ai/contact.
There are two ways to get started. Organisations can upload footage directly to the Vision Platform at vision.reka.ai and run natural language queries against their own content with no sales conversation required. For organisations ready to scope a pilot or discuss deployment options for specific infrastructure requirements, a demo can be requested at reka.ai/contact.
Reka provides API access to its full model family through Reka Managed Cloud. Developers and enterprises can access Reka Core, Reka Flash, Edge, and Spark models via a standard API. Reka models are also available through Oracle Cloud Infrastructure (OCI) for organisations operating within the Oracle ecosystem.
Reka provides API access to its full model family through Reka Managed Cloud. Developers and enterprises can access Reka Core, Reka Flash, Edge, and Spark models via a standard API. Reka models are also available through Oracle Cloud Infrastructure (OCI) for organisations operating within the Oracle ecosystem.