Evidence intake, organizer, and timeline
Upload the record, index it for RAG, organize evidence into themes, and build a chronological case narrative.
SideBar turns your case record into sourced research, judge and counsel profiles, Motion Lab scenarios, adversarial simulations, Jury Swarm deliberations, reusable workflows, voice read-aloud, and final strategy reports.
Each stage persists to the case file as soon as an agent finishes, so the team can move between research, simulation, motions, workflows, and reports without losing context.
Upload documents, organize facts, index evidence for RAG, and arrange the timeline that grounds every later agent.
Research Swarm builds judge, counsel, and issue memos with source ledgers instead of unsourced AI summaries.
Motion Lab applies jurisdiction-first research to simulate briefing, hearing pressure, predicted rulings, and next actions.
Dual-model advocates argue from the record while a judge model scores each round and preserves evidence and authority citations.
Reusable workflow diagrams run agents and tools over case context, with voice read-back for long prompts and outputs.
Reports collect the strategy, simulations, vulnerabilities, citations, preparation tasks, and next steps into one case artifact.
SideBar now covers the preparation loop from intake and evidence through motion practice, simulations, jury testing, workflows, voice read-aloud, and final strategy reporting.
Upload the record, index it for RAG, organize evidence into themes, and build a chronological case narrative.
Research Swarm gathers source ledgers before synthesis, so profiles and memos stay auditable.
Test motions before filing: rules, local-practice risks, briefing simulation, hearing questions, and predicted rulings.
Codex and GPT argue opposing sides while a separate judge model scores each round and records citations.
Seat a synthetic jury, run deliberation, inspect votes and rationales, then feed the result into the report.
Run reusable legal workflows, draft documents in chat canvas, and read long arguments aloud with ElevenLabs voices.
The web edition is built for teams, shared cases, and cloud-hosted AI/research workflows. The macOS edition remains available for a local-first desktop workflow with native SwiftUI ergonomics.
Supabase Auth, row-level isolation, case collaboration, Research Swarm, Motion Lab, Jury Swarm, Workflows, and reports.
The macOS build remains available for attorneys who want the native SideBar app alongside the hosted workspace.
Cases, evidence, profiles, and reports are scoped to your firm by Postgres row-level security. Models never see another firm's record. Your work is exportable on demand.
Every case, every document, every report is bound to your org by Supabase RLS. The keys to your data sit with your org owner — bring your own Supabase project if you need full data sovereignty.
TLS in transit. Postgres column encryption for sensitive fields. Storage objects served behind short-lived signed URLs scoped to the requesting member.
Cases live as standard Postgres rows; reports as Markdown. Export to JSON or PDF in one click. When you delete a case, the rows and the Storage blobs are both gone — no soft-delete shadow.
Screen-by-screen specs, the Supabase schema, the Edge Function surface, the RAG pipeline, agent prompts, and the streaming protocol — all open in the repo.