CoreMF · Core Media Framework

One framework for RTC, AI, GPU, media processing, transcoding and playback.

Combine live streams, software/hardware codecs and GPU filtering. Access and process individual video frames and audio samples through custom multi-language plugins for AI, processing and analysis. Configure graphs in code or JSON and deploy on your infrastructure.

Custom video and audio processing uses the matching plugin or source-development interface. Choose your integration path ↗

CoreMF’s advantages

RTC, plugin access to frames and samples, GPU processing and flexible encoding—integrated in one native SDK.

One engine for live and offline workflows
Reuse RTC, streaming, file processing, recording and rendering components.
Media gateway construction
Connect RTMP/RTSP/SRT ingest to RTC/SFU, recording or streaming output.
Programming and JSON configuration
Build graphs in code or JSON; reuse templates and subgraphs.
GPU processing and metadata-driven composition
Combine CoreGL shaders, effects and layouts with metadata-driven graphics.
Production and editing foundations
Build on timelines, scenes, previews and render jobs.
Video frame and audio sample access through multi-language plugins
Access individual video frames and audio samples through native, Python or JavaScript plugins. Packet access follows the matching interface.
Flexible encoder selection
Auto-select eligible encoders; choose hardware or an explicit customer BYO executable.
Component reuse
Adapt selected native classes under agreed source or consulting terms.
Compact application SDK boundary
Integrate binary applications through the single CoreMF_API.h public header.
Protection integration points
StreamCrypto and DRM packaging hooks; key/license authority and provider qualification remain separate.

Application languages: C/C++, Python, Node.js, Java and Go · Platform builds: Linux/Ubuntu, macOS and Windows.

Actual engine output · 29 Oct 2026

See actual input and output.

See real footage processed with CoreGL color grading, custom shaders and animated composition inside the media pipeline.

InputBigora waterfall: fourteen seconds of real footage, 1280×720 at 25 fps, with recorded audio.
CoreGL GPU productionColor, vignette, particles, titles and an animated detail view through runAPI. The 13.96-second output and its audio fully decode.

Footage: “Bigora waterfall” by Termininja, CC0. Trimmed and resized for this demonstration; the inset is a crop of the same recording.

Windows development build; Intel UHD 770. Full output decode passed. This is a functional demonstration, not a commercial package or latency/capacity benchmark.

Recorded RTC verification

Local sessions under sustained load

Four publishers and four viewers in a one-hour local CPU RTC/SFU fixture. The recorded profile covers frame rate, latency, memory and reconnect behavior.

Read the test scope and results ↗

Recorded SDK verification

Processing through an installed plugin

A native ABI template compiled against the source SDK and processed real video through the installed API runner. This evidence is specific to that native video path.

Inspect the integration evidence ↗

Adopt it your way

Choose your delivery and integration path.

Embed

CoreMF Media SDK

Integrate through one public header; configure graphs in code or JSON. Choose binary or agreed source delivery.

Explore the SDK

Customize

Integration & customization

Commission component customization, product integration or performance and deployment engineering for your application.

Explore services

Build with CoreMF

Tell us what your product needs to do.

Tell us your project, target platform and main requirement.

Discuss your project