All projects matching “chunking”
3 results- turboshop.aiService
High-speed single-pass state machine microservice that deterministically repairs cut-off and truncated JSON streams from LLMs (e.g., max_tokens cut-offs or live SSE chunks). Automatically closes unterminated strings, completes partial literals (tru/fals/nul) and numbers, resolves dangling keys/colons, strips markdown fences, and balances nested containers in LIFO stack order with optional progress
Baseturboshop.aifrom $0.0124 endpoints39calls / 30d - Transcript PackService
Sells a one-time Transcript Pack job token for audio up to 10 min ($0.10 USDC on Base). Use the token with POST /api/job, upload 16 kHz mono 16-bit WAV chunks (max 60 s each) to /api/job/{id}/chunk/{i}, then POST /api/job/{id}/finish to get a transcript (.txt), captions (.vtt) and AI show notes (.md). See /llms.txt.
Basetranscript-pack.toolschemaforge.workers.devfrom $0.101 endpoint0calls / 30d - Melchior Video ContextService
Find transcript passages matching a phrase or terms in a public YouTube video, with timestamp citations.
Basetranscript.melchiorlabs.comfrom $0.0033 endpoints0calls / 30d