This roadmap communicates direction, not dates or delivery commitments. Each item still requires a scoped issue, review, and the quality gates in CONTRIBUTING.md.
- Complete community application: Python engine, FastAPI service, background worker, Next.js workspace, PostgreSQL migrations, and reference Caddy/Docker Compose deployment.
- Literature discovery, library and review, projects, Data Lab, figures, surveys, text and spoken interviews, knowledge, brainstorming, and manuscript workflows.
- Operator-supplied mail, scholarly-metadata, model, speech, search, and parser integrations with explicit configuration and processing gates.
- Fail-closed local/TLS deployment configuration, bounded resources, backup and restore tooling, and authenticated erasure-journal replay.
- Engine, API, web, contract, self-hosting, security, dependency, container, DCO, and release automation.
- Apache-2.0 licensing with separate trademark terms. Billing, commercial plans, the hosted production deployment, and the marketing site remain separate.
- Version the API and event-stream contracts with machine-readable compatibility fixtures.
- Add synthetic browser end-to-end tests for onboarding, representative research workflows, public participation, retention, and account erasure.
- Publish migration rehearsal fixtures for every supported upgrade path.
- Define portable provenance envelopes for project inputs, provider settings, transformations, revisions, and output hashes.
- Expand deterministic research exchange formats, beginning with bounded RIS/BibTeX import and export with explicit loss reporting.
- Harden accessibility, CSP, browser storage, rate limits, and capability-link handling against an explicit threat model.
- Test encrypted off-site backup and isolated restore recipes against released images on a schedule.
- Publish multi-architecture container images only after both architectures pass the same smoke and migration gates.
- Add portable object storage without weakening tenant isolation, erasure, or backup consistency.
- Add deployment observability with privacy-preserving defaults and no request bodies or public capability URLs in logs.
- Define a supported high-availability profile after measuring database, storage, worker, and provider failure modes.
- Maintain adversarial provider evaluations for spoken and generated research workflows without storing participant content in public evidence.
- Add cross-database query translation conformance fixtures that surface syntax loss rather than hiding it.
- Improve OpenAlex paging, retry, timeout, provenance, and redaction behavior.
- Extend source-anchored screening, extraction, synthesis, and manuscript workflows with deterministic conflict behavior.
- Add reproducible scientific visuals and statistical recipes only with explicit assumptions, validation fixtures, and method limits.
- Improve interchange for consented survey and interview evidence using synthetic public fixtures.
- Payments, pricing, subscriptions, checkout, and commercial-plan enforcement.
- Hosted-service administration, private production configuration, customer data, production evidence, analytics, and provider credentials.
- The SixSentences marketing and landing website.
- Redistribution of proprietary corpora or scholarly full text without rights.
- Any claim that software alone guarantees completeness, compliance, causality, or scientific validity.
Track implementation through GitHub issues and use discussions for ideas that are not yet ready to scope.