refray/README.md

refray

A tool to keep your repos in sync across all git platforms, while being able to work from everywhere all at once.

Created becasue github is so unusable and unreliable and I want to leave, but I don't want to leave the community behind.

• ∞-side sync: Sync between any number of hosted/self-hosted git accounts/orgs/groups
• read-write mirrors: Make changes from any provider, and the changes will sync to the others
• webhook support: Sync right after push, reduce potential divergence window
• conflict handling: Rebase or open pull requests when two platforms diverge
• tracks deletions: Delete branches/repos across platforms when they are deleted from one platform
• selective sync: Sync subset of repos by regex white/black list, or by private/public visibility

Supported platforms: GitHub, GitLab, Gitea, Forgejo

Install

Option 1. Install from source

1. Install rust cargo if you don't have it: https://rustup.rs
2. cargo install refray

Option 2. Download binary

Go to the releases page, find the latest release, and download the appropriate binary for your platform.

Option 3. Docker Compose

docker compose run --rm refray config

# Start the webhook receiver as a service
docker compose up -d --build

# If you want to edit config manually:
docker compose run --rm --entrypoint nano refray /data/config/refray/config.toml

Configure

Run the interactive configuration wizard:

refray config

One-time Sync

Run all configured mirror groups:

refray sync

Sync options

Run one group:

refray sync --group personal

Preview commands without writing to Git remotes:

refray sync --dry-run

Sync only repositories whose names match a regex:

refray sync --repo-pattern '^(foo|bar)-'

For persistent per-group filters, set repo_whitelist and/or repo_blacklist in the config instead. --repo-pattern is an extra one-off filter applied on top of the group config.

Retry only repositories that failed during the previous non-dry-run sync:

refray sync --retry-failed

Control repo-level parallelism:

refray sync --jobs 8

Service & Webhooks

You can run refray as a service that listens for webhook events and runs full sync periodically. This is the recommended way to run refray.

If you want to use webhooks, you need to expose port 8787 to a public URL that can be accessed by the git provider (e.g. using port forwarding, reverse proxy, or cloudflare tunnel).

Start the receiver:

refray serve

Expose that listener with your reverse proxy or tunnel, then install repository webhooks. If [webhook] is configured, the URL and secret can come from config:

refray webhook install

Useful install filters:

refray webhook install \
  --url https://mirror.example.com/webhook \
  --secret-env REFRAY_WEBHOOK_SECRET \
  --group personal \
  --repo-pattern '^important-'

The receiver accepts POST / and POST /webhook. It verifies GitHub/Gitea HMAC SHA-256 signatures and GitLab webhook tokens, then queues refray sync --group <group> --repo-pattern '^<repo>$' internally. Duplicate events for the same group/repo are coalesced while a job is queued or running. Sync jobs are serialized inside the receiver so the local ref and failure caches stay consistent.

When [webhook].install = true, normal refray sync also checks webhook installation status and installs missing webhooks for repositories that have not been recorded yet. Installation status is stored in webhook-state.toml under the work directory.

To uninstall webhooks previously installed by refray:

refray webhook uninstall

Manual webhook uninstall checks repositories on the provider instead of trusting only local state. To uninstall one repository exactly:

refray webhook uninstall important-repo

To move installed hooks to a new public URL, use webhook update. It removes hooks matching the current configured [webhook].url, installs the new URL, updates [webhook].url in the config, and refreshes local webhook state:

refray webhook update --url https://new.example.com/webhook

Serve can also run periodic full syncs. The interval can be configured in [webhook].full_sync_interval_minutes or overridden at startup:

refray serve --full-sync-interval-minutes 30

If [webhook].reachability_check_interval_minutes is configured, serve periodically checks that the public webhook URL is still reachable and logs a warning when it is not.

Sync Semantics

Each mirror group is treated as a set of equivalent namespaces. Repositories are matched by repository name across all endpoints.

Set sync_visibility = "all", "private", or "public" on a mirror group to choose which repository visibility is included in that group. visibility still controls the visibility used when refray creates a missing repository.

Set repo_whitelist = ["..."] and/or repo_blacklist = ["..."] on a mirror group to filter repository names with regular expressions. An empty whitelist includes all repository names, and blacklist matches are excluded after whitelist matches. These name filters are independent from sync_visibility; both must match for a repository to be synced.

For every repository name found in any endpoint, refray will:

1. Create missing repositories on the other endpoints when create_missing = true.
2. Fetch all branches and tags from each existing endpoint into a local bare mirror cache.
3. Compare branch tips across endpoints.
4. Push the winning branch tip to every endpoint.

Branch conflict handling is intentionally conservative:

• If all endpoints agree on a branch tip, that tip is pushed everywhere.
• If one branch tip is a descendant of the others, the descendant wins and is pushed everywhere.
• If branch tips diverged, conflict_resolution controls what happens.
• If allow_force = true or refray sync --force is used, a diverged branch chooses the newest commit timestamp and force-pushes it.

Conflict resolution strategies are configured per mirror group:

• fail: fail the repository sync when branch tips diverge.
• auto_rebase: rebase divergent commits in endpoint order into one branch history, push fast-forward updates normally, and force-push only endpoints whose original tip was rewritten. If rebase hits a file conflict, fail.
• pull_request: push temporary refray/conflicts/... branches and open provider pull requests/merge requests so a person can resolve the divergence.
• auto_rebase_pull_request: try auto_rebase first, then fall back to pull requests if rebase hits a conflict.

When a previously opened conflict pull request is merged, the next sync sees the merged branch as the winning tip, pushes it to the other endpoints, and closes stale refray/conflicts/... pull requests for that branch.

Repository and branch deletion are propagated only when it is safe to infer intent. If a repository existed on every endpoint in the previous successful sync, then disappears from one endpoint while the remaining endpoints still have the previous synced refs, refray deletes it from the remaining endpoints instead of recreating it. If the repository was deleted everywhere, refray removes its saved sync state. If the repository was deleted on one endpoint but changed elsewhere, it is treated as a conflict and skipped.

Branch deletion follows the same rule at branch scope: if a branch existed on every endpoint in the previous successful sync, then disappears from one endpoint while the remaining endpoints still have the previous tip, refray deletes it from the remaining endpoints instead of recreating it. If the branch was deleted on one endpoint but changed elsewhere, it is treated as a conflict and skipped.

Tags are fetched into provider-specific cache refs and pushed only when the tag object agrees across providers or exists on one side. Divergent tags are skipped and reported. Tag deletion is not propagated.

Example Config

[[sites]]
name = "github"
provider = "github"
base_url = "https://github.com"
token = { env = "GITHUB_TOKEN" }

[[sites]]
name = "gitea"
provider = "gitea"
base_url = "https://gitea.example.com"
token = { env = "GITEA_TOKEN" }

[[mirrors]]
name = "personal"
sync_visibility = "all"
repo_whitelist = ["^important-"]
repo_blacklist = ["-archive$"]
create_missing = true
visibility = "private"
allow_force = false
conflict_resolution = "auto_rebase_pull_request"

[[mirrors.endpoints]]
site = "github"
kind = "user"
namespace = "hykilpikonna"

[[mirrors.endpoints]]
site = "gitea"
kind = "user"
namespace = "azalea"

Testing

Run the normal, non-destructive test suite:

cargo test

The sequential live e2e test is ignored by default because it creates and deletes repositories on real provider accounts. Configure four token/username pairs in .env or the process environment:

GH_USER=...
GH_TOKEN=...
GL_USER=...
GL_TOKEN=...
GT_USER=...
GT_TOKEN=...
FO_USER=...
FO_TOKEN=...

Optional base URL overrides are GL_BASE_URL, GT_BASE_URL or GITEA_BASE_URL, and FO_BASE_URL or FORGEJO_BASE_URL. The Gitea and Forgejo defaults are https://gitea.com and https://codeberg.org.

Run the destructive e2e test explicitly:

REFRAY_E2E_ALLOW_DESTRUCTIVE=1 \
  cargo test --test sequential -- --ignored --test-threads=1 --nocapture

By default cleanup only deletes repositories named refray-e2e-*. To start by deleting every owned repository visible to the configured accounts, set REFRAY_E2E_CLEAR_ALL_REPOS=DELETE_ALL_OWNED_REPOS. Provider skips (REFRAY_E2E_SKIP_GITHUB, REFRAY_E2E_SKIP_GITLAB, REFRAY_E2E_SKIP_GITEA, REFRAY_E2E_SKIP_FORGEJO) and REFRAY_E2E_ALLOW_PARTIAL=1 are available for local debugging, but the full support check should run with all four providers.

Issues and Pull Requests

Issues and pull requests are not mirrored.