Skip to main content
Teleox.ai

Systems

Systems prove the framework has implementation surfaces.

Context Graph, ClipCannon, OCR Provenance, and Dynamic / ME-JEPA are witnesses around the same source-grounded posture.

Context Graph diagram with 13 embedders and pairwise interactions.
Context Graph diagram with 13 embedders and pairwise interactions.

Figure 1.Context Graph is the text-side DDA witness: N=13, up to 91 signals per input as a constructive upper bound.

Figure description

The figure represents Context Graph as a text-side DDA witness with 13 frozen embedders. The direct projections contribute 13 signals per input and the pairwise interactions contribute up to 78 more, for up to 91 structured signals per input as a constructive upper bound pending mutual-information audit.

ClipCannon pipeline diagram showing staged analysis from source video to multimodal records.

Figure 2.ClipCannon materializes video into multi-modal records through a 23-stage analysis DAG.

Figure description

The figure shows source video moving through a staged ClipCannon DAG. Analysis stages extract visual, semantic, emotion, speaker, prosody, sentiment, and voice-identity records. The outputs are persisted as multi-modal training and provenance records rather than one unstructured video blob.

Document provenance diagram showing source files, OCR, chunks, embeddings, and hash-linked records.

Figure 3.OCR Provenance is the document-side witness for local-first, hash-linked transformation history.

Figure description

The figure traces a document from source file through OCR, chunking, embedding, and retrieval records. Each transformation writes provenance metadata and hash-linked state, so later consumers can inspect where a derived text segment or embedding came from.

Production witnesses

Each system page explains what construct the system witnesses and which limitation travels with that claim.

text-dda-witness

Context Graph

Text-side DDA witness using 13 frozen embedders, RocksDB storage, and MCP retrieval tools.

Embedder panel
N=13
Structured signals
up to 91 per input

Open system

video-dda-witness

ClipCannon

Video-side DDA/TCT witness using a 23-stage DAG, seven modalities, and per-project provenance records.

Pipeline
23-stage DAG
Modality panel
N=7

Open system

document-provenance-witness

OCR Provenance

Document-side provenance witness for local-first OCR, chunking, embeddings, and hash-linked records.

Architecture
local-first MCP server
Provenance
hash-linked transformation records

Open system

jepa-runtime-witness

Dynamic / ME-JEPA

Private-preview JEPA-style runtime witness for audited bundles, deterministic artifacts, and Full State Verification.

Preview artifact
private PDF
Verification posture
source-of-truth readback

Open system

Sources

  • docs2/PAPER.md#4
  • docs2/alldata.md