A tinkering programmer and author of "Spring Cloud Microservices in Action"
Using Doubao Work's Infinite Canvas, a non-artist converts a 70-chapter novel into a consistent vertical webtoon, with an AI agent handling storyboarding, panel generation, self-correction, and interactive dialogue fixes across two chapters.
OctaFuse Gateway 2.14.0 introduces OpenAI-compatible adapters for MiniMax, Seedream, and Gemini image and voice models, adds per-route-group user discounts with price preview, and switches admin config to per-field saves to prevent overwriting live usage data.
The author experiments with Doubao Work AI agent to simulate a first-time user testing their AI writing desktop app "Lingxie Creation", having it explore the UI, write a novel, and produce a prioritized experience report with issues and optimization suggestions, then proposes integrating this into QA workflows as reusable Skills and a QA manager role.
OctaFuse Gateway 2.13.0 introduces three major upgrades: supplier discounts now support time-bound validity and multi-rate pricing separating public promotions from user billing and actual costs; route editing gains instant pre-save testing with upstream request preview; and usage analytics splits into overview, cost/usage, and performance views with refined metrics like weighted success rate and failover request ratio.
OctaFuse Gateway 2.12.0 adds supplier account categorization with live quota queries, a searchable routing workspace showing sticky binding counts, multimodal model endpoint discovery via /v1/models with protocol and operation fields, responsive mobile admin UI, and extends streaming idle timeout from 30 to 90 seconds.
OctaFuse Gateway 2.9.0 introduces per-user and per-key RPM rate limiting with sliding 60-second windows, per-route custom HTTP headers for upstream providers, a redesigned admin dashboard with consolidated metrics, and Key-level analytics with ingress host tracking, plus migration 0028 for database schema updates.
A developer shares three practical automations using Doubao Work Agent: aggregating user feedback from Jindata forms and GitHub Issues, generating weekly reports from GitHub commits and PRs, and monitoring AI model pricing via web parsing, all orchestrated through connectors, scheduled tasks, and reusable Skills.
OctaFuse Gateway 2.6.0 adds project‑level Vertex AI service‑account authentication, introduces token‑ and per‑image billing modes for image models, refines time‑slot pricing with override mode, and upgrades the admin UI, Playground and Simulator with new provider templates and routing visualizations, while clarifying failure‑request charge rules.
Based on real‑world request logs collected via octafuse‑gateway, the article compares cache hit rates and availability of major Chinese LLM vendors, showing that official model providers (e.g., DeepSeek, Xiaomi MiMo, Zhipu) achieve over 90 % hit rates, while cloud MaaS and Volcano Ark lag behind, especially in high‑frequency Agent scenarios.
The article analyses how EdgeOne Makers provides a comprehensive engineering foundation—including managed runtime, sandbox tools, conversation storage, observability, model integration, authentication, and Git‑based deployment—to turn AI Agent demos into production‑ready services, and compares it with alternative approaches.
