I design tools that help agents and humans keep context, recover threads, and turn messy daily inputs into durable workflows.
Explore Projects / Agent Skills / Telegram Scanner / Codeksei
|
Companion agents Local-first assistants that can keep timelines, reminders, reviews, diaries, and project context close to the work. |
Agent memory & continuity Workflows for recalling useful context, resuming long-running threads, and leaving durable state across tools. |
Signal extraction tools Practical scanners and report generators that turn noisy channels into LLM-ready summaries and decisions. |
|
codeksei Local-first companion agent infrastructure for memory, timeline, reminders, reviews, and project continuity. companion agents / local-first / automation
|
tg-channel-scanner Scan Telegram channels, filter signals with Markdown profiles, and generate LLM-ready reports. signal extraction / reports / LLM workflows
|
|
my-agent-skills Internal personal agent skills for Codex, Claude Code, and Hermes workflows. skills / workflow assets / agent practice
|
CLI-Agent-Orchestration Internal CLI worker manager for multi-agent dispatch, tracking, cost, and quality loops. multi-agent / CLI / quality loops
|
I am working toward local-first AI companions, agent-native workflows, cross-tool memory, and human-in-the-loop automation that stays useful across long-running work.
I care about tools that reduce context loss, especially for long-running projects and attention-fragmented work. The best systems leave a trail you can return to: what happened, what matters now, and where to restart.