Digital quest archiving with privacy and chronological indexing. The foundational binding structure of EvoBioSys — tools and infrastructure that enable flow-based working. Open-source, local-first, designed to liberate rather than lock in.
etymology
kidur = foundational binding structure
kidur is not a product. It is the substrate on which other holons build. Where other holons ask what to grow or who to serve, kidur asks a different question: what tools enable people to do their deepest work without friction?
purpose
Create a singular structure for yourself first, then offer it to others — a space where anyone can upload their quests, including legacy quests from people who have died without fulfilling their mission. Privacy is essential. This is an extended archival function.
Giving ourselves the same access as billionaires — a shared brain.
The answer is a constellation of projects unified by a single concept: FloWork.ing — flow-state working — where the technology disappears and only the creative process remains. Tools that enable FloWork.ing are tools that protect attention, align time with energy, and keep creative work uninterrupted.
Voice message processing via local LLM. Speak your thoughts, kidur transcribes and indexes them — no cloud required.
Message directly from within kidur. Communication without context-switching, without leaving your flow state.
Multilingual from the ground up. Your quests and knowledge should cross language barriers as easily as they cross tool barriers.
Intelligence woven into the infrastructure layer, not bolted on. AI as a multiplier for building and maintaining the commons.
Sub-node level collaboration via CRDTs. Multiple people editing the same structure simultaneously, conflict-free.
Not privacy by policy. Local-first storage, end-to-end encryption, and data sovereignty are structural, not optional.
stack
# kidur — Tech Foundation Stack # The foundational binding structure of EvoBioSys [core] name = "kidur" etymology = "ki (place/base/earth) + dur (bond/tie/enclosure)" purpose = "Digital quest archiving with privacy and chronological indexing" paradigm = "local-first, open-source, flow-optimised" [database] engine = "Apache AGE" query_lang = "Cypher + SQL" base = "PostgreSQL" note = "Graph queries over relational storage — best of both worlds" [collaboration] engine = "Yrs (y-crdt)" granularity = "sub-node" sync = "WebSocket" note = "Conflict-free replicated data types for real-time collaboration" [ai] inference = "local LLM (voice, translation, indexing)" backend = "AI-driven pipeline" philosophy = "intelligence in the infrastructure, not bolted on" [features] voice = "local LLM transcription and processing" messaging = "direct messaging from within kidur" translations = "multilingual by default" privacy = "local-first, end-to-end, structural not optional" archival = "chronological indexing, legacy quest support" [principles] foss = true local_first = true flow_over_features = true interop_over_integration = true
The choice of Apache AGE is deliberate: it lets kidur run Cypher graph queries on top of PostgreSQL, combining the expressive power of a graph database with the reliability and ecosystem of Postgres. No separate graph database to maintain, no data duplication — just one engine that handles both relational and graph queries.
Yrs (the Rust implementation of Y.js CRDTs) provides sub-node granularity for collaboration. Two people can edit the same paragraph simultaneously, and the data structure resolves conflicts automatically. No central server required for merge resolution.
ls ./sub-holons/
Each sub-holon addresses a distinct layer of the flow-working stack — from knowledge capture to scheduling to distraction-free creation.
A note-operating system designed to interoperate across knowledge silos. Local-first, FOSS, built for people who think in multiple tools and need them to talk to each other.
A category of flow-state tools — software that protects attention, aligns time with energy, and keeps creative work uninterrupted.
A text weaving tool for distraction-free, structured writing. Designed for authors and thinkers who work in plain text and need composability over complexity.
The connective tissue — self-hosted services that keep the stack sovereign and interoperable. The plumbing nobody sees but everybody depends on.
cat principles.md
kidur is built on a small set of non-negotiable principles that shape every tool in the constellation:
FOSS by default. Every tool is free and open-source wherever possible. What can be built by others should be — then copied, forked, and improved through the open-source commons.
Local-first. Data lives on your device. Cloud sync is opt-in, never required. Sovereignty starts with where your files sleep at night.
Interoperability over integration. Rather than building one monolithic platform, kidur creates bridges between existing tools — respecting that people already have workflows they trust.
Flow over features. Every design decision is filtered through one question: does this protect the user's flow state, or interrupt it?
AI as multiplier. Much of kidur's roadmap can be accelerated through AI-assisted development — building tools faster, then releasing them into the commons for others to extend.
The best tool is the one you forget you are using.
kidur exists because the current tooling landscape forces a choice between power and freedom, between capability and sovereignty. The answer is not to compete with Big Tech feature-for-feature, but to build the minimum viable infrastructure for humans to do their best work — and then get out of the way.
query --semantic-web
kidur also encompasses work on the semantic web and up-to-date, structured knowledge systems. The goal is to make information not just searchable but meaningful — machine-readable, context-aware, and connected across domains.
This layer bridges the gap between kidur's local-first tools and the broader web, ensuring that knowledge created in the EvoBioSys ecosystem can participate in the open, linked-data commons. Apache AGE's graph capabilities make this a natural extension — the same engine that powers quest archiving can serve RDF-like queries over structured knowledge.
-- Example: finding connected quests across the knowledge graph MATCH (q:Quest)-[:RELATES_TO]->(topic:Topic) WHERE topic.domain = 'regenerative-agriculture' RETURN q.title, q.author, q.status, topic.name ORDER BY q.created_at DESC;