VQV Signal

AI AT WORK WATCH TECHNICAL

SB Intuitions publishes Sarashina guides for coding-agent setups

According to a post on X by @sbintuitions, SB Intuitions, a SoftBank group company, published a tech blog on connecting its Sarashina APIs to coding agents. The guide covers using OpenCode as the main model, Japanese-specialized subagents, and delegating tasks from Claude Code. It also includes configuration examples for Hermes Agent and Crush. The source describes these as vendor recipes for using a regional LLM inside existing agent tools.

Source: Scobleizer News Pulse (X) · x.com Found via Robert Scoble’s Scobleizer News Pulse · a post on X, not independently verified Published 2026-11-06T01:40:00+00:00 Detected 2026-11-06T18:21:04+00:00
View original source

According to a post on X by @sbintuitions, SB Intuitions, a SoftBank group company, published a tech blog on connecting its Sarashina APIs to coding agents. The guide covers using OpenCode as the main model, Japanese-specialized subagents, and delegating tasks from Claude Code. It also includes configuration examples for Hermes Agent and Crush. The source describes these as vendor recipes for using a regional LLM inside existing agent tools.

AI-assisted summary based on the listed source.

SoftBank's SB Intuitions published a tech blog on wiring Sarashina APIs into coding agents: using OpenCode as the main model, Japanese-specialized subagents, and delegating tasks from Claude Code, with Hermes Agent and Crush configuration examples. Why it matters: Vendor recipes for Claude Code and OpenCode make a...

Developers who already use Claude Code or OpenCode and want a Japanese-focused model now have written setup examples to try. Teams considering Sarashina for coding workflows can compare these configurations with their current tools. Watch whether the examples are kept up to date and whether other agent tools receive similar guides.

Teams using AI at work may want to compare this against current productivity and review workflows.

Signal Strength 90% Technical label WATCH Public Interest 47 Category AI AT WORK Reader Depth TECHNICAL

Signal Strength reflects source quality, relevance, freshness and evidence. Public Interest helps organize discovery; it is not proof of truth.

Public Interest components
Recognizable Entity Score 51 Practical Impact Score 28 Novelty Interest Score 70 Consequence Score 46 Curiosity Score 16 Shareability Score 57

VQV surfaced this signal because it is recent, relevant to AI Coding Tools, connected to Scobleizer News Pulse (X).