Yes. The README directly supports imports from these sources and full JSON export, although the result depends on how the source fields map.

• The Accounts and data section lists imports from FitNotes, Strong, and Hevy; FitNotes-style imports use CSV, while Strong and Hevy can also use an API key. Apple Health weight exports are supported too.
• The same section says all data can be exported as one JSON file whenever needed, matching the requirement for portability.
• The Progress section allows saved workouts to be edited, paper workouts to be added, and dates to be corrected; records are re-read from the corrected history.
• The README does not promise identical handling of every source app’s fields, units, exercise names, or date formats, so migration accuracy is not fully guaranteed.

No, not as a full organizational fitness-management system. It can support a small multi-profile self-hosted instance, but the README does not promise coach approval, team permissions, or high-concurrency operation.

• The README provides a separate Kubernetes self-hosting guide, so the deployment form is supported.
• The configuration reference includes INVITE_ONLY, ADMIN_UIDS, and AUDIT_LOG, while Accounts and data lists an optional admin dashboard.
• However, the project insights state that JSON-file storage is suited to personal or small-scale instances, not large user counts, high concurrency, multi-tenancy, or complex queries.
• The same analysis says the project targets personal fitness tracking and lacks medical governance, coach approval workflows, and organization-level permissions.

A household or small private deployment may fit, but formal team management exceeds the documented security and data model.

It depends: the base deployment fits an ARM64 NAS, but remote phone use with Passkeys depends on correct HTTPS and domain configuration.

• The README explicitly provides prebuilt amd64 and arm64 images and says Node.js is not required on the host.
• Data is kept in a folder controlled by the user, which suits NAS persistence and backups.
• Phone Passkeys require HTTPS on a domain, with matching RP_ID and ORIGIN; running Compose alone is not sufficient.
• The first start downloads about 140 MB of exercise media, so the NAS needs enough storage and bandwidth.

It is straightforward on a LAN, but remote use requires a reverse proxy or a service such as Cloudflare Tunnel.

Yes. It fits because it combines routines, effort ratings, progression, and muscle status instead of storing only weight and reps.

• The Planning section includes Push/Pull/Legs, Upper/Lower, Full Body, and 5×5 starter plans, all editable as normal routines.
• The Training section supports optional RIR or RPE, prefilled previous weights, PR detection, supersets, warm-ups, and drop sets.
• The Progress section provides linear progression, Greyskull LP, double progression, and time progression; missed reps do not add load, and stalls trigger a deload.
• It also provides estimated 1RM, training-volume views, recovering muscles, and untrained muscles, matching long-term progression tracking.

The limitation is that these indicators depend on exercise classification and accurate logging; they are not medical or precise physiological measurements.

Yes, but these are two optional features with different deployment boundaries, and MCP is not part of the core Docker build.

• The Optional extras section says the AI coach runs on the server with the user’s own provider key and supports Anthropic, OpenAI, Gemini, OpenAI-compatible endpoints, and Ollama.
• The coach drafts a training week or suggests changes, while the README explicitly requires user approval for every change.
• The MCP server is intended for Claude Desktop and is read-only and local, matching a history-query requirement.
• The README also states that MCP is not part of the Docker build, so it requires separate setup from its own documentation; docker compose up -d should not be assumed to enable it.

This fits users who want controlled AI assistance rather than automatic plan mutation.

It depends. The README explicitly supports offline use, multi-device sync, and merging concurrent edits, but it does not describe a full version-control system.

• The Why openGym section says the app is installable on the home screen, works offline, and syncs between phone and laptop.
• The Accounts and data section explicitly says edits from two devices merge instead of overwriting each other.
• The workout interface also keeps the screen awake and provides rest timers and alerts, fitting gym-phone use.
• However, the project insights note that complex conflicts may still require review. The supplied README excerpt does not define conflict precedence, maximum offline duration, or a history-recovery interface.

It is suitable for normal workout logging, but simultaneous edits to the same complex routine should not be assumed to resolve perfectly.