Filenergy

A multi-tenant private file host that turns into a searchable, askable knowledge base. Drop PDFs, DOCX, Markdown or text in and Filenergy extracts the text, embeds it, and lets you (or your team) chat with the library — with citations, a programmatic API, and Stripe billing built in.

Stack: Flask 3 + SQLite, Voyage embeddings, the Anthropic Claude API for answers, and Stripe Checkout for billing.

Screenshots

Chat with citations + vision attachment

File library with bulk multi-select

Answer-quality dashboard (thumbs feedback over time + triage queue)

What's new in this release

Reranking on top of vector search — every retrieval pulls a wider candidate set, then a fast Claude pass scores each chunk 0–10 against the query. Disable with FILENERGY_RERANKER=noop.
SCIM 2.0 at /scim/v2/* — provision and de-provision users from Okta, Workspace, Auth0, Azure AD. Bearer-token auth + X-Filenergy-Workspace-Slug header.
Email-to-ingest — every workspace gets a deterministic inbox-<slug>-<token>@<domain> address. Forward an email and the body + attachments become indexed files.
Soft-delete with undo — bulk-delete shows a toast with an Undo button for 8 seconds. Files only get hard-deleted after a grace window (flask purge-deleted-files runs from cron).
Conversation pinning + archiving — pinned threads float to the top of the sidebar; archived threads are hidden but kept for export.
File rename, collection share links, command-palette arrow-key navigation, redesigned landing page.

UX

Command palette (⌘K / Ctrl+K / /) — quick-jump to any page, searchable, keyboard-driven.
Chat power moves — Stop generation mid-stream, Regenerate the last answer, thumbs up/down per assistant message (feeds the eval signal), copy any message, inline rename of conversations, deep-linkable starter questions.
File library — bulk multi-select with checkbox column, bulk delete + reindex, instant client-side filter, drag-drop upload zone with per-file progress and inline retry.
Mobile-first nav — hamburger drawer on small screens; chat & settings layouts collapse cleanly on mobile.
Toasts — flash messages render as dismissable toasts (auto-hide on success, manual dismiss on error). window.fnToast(msg, kind) is available everywhere for ad-hoc notifications.
SVG icon sprite baked into base.html — every nav item, button, and empty state uses a consistent line-icon set.

Highlights

Workspaces (multi-tenant) — every user gets a personal workspace and can be invited to others. Roles (owner / admin / member). Members, invitations, and a switcher live under /settings/workspace.
Collections — group files into folders ("notebooks") and ask questions scoped to one collection or one file. Per-doc Q&A like NotebookLM / Humata.
Auto-summary + suggested questions — on first index every file gets a one-line summary and 3 suggested questions, generated by Claude with structured output (output_config.format). The "magic moment" of the demo.
Streaming chat — Server-Sent Events feed tokens to the browser as Claude generates them. Markdown is rendered live; sources cite back to the originals. Multi-turn threads persist with the last 12 messages of context sent back to the model. Supports collection_id / file_id scoping.
Background indexing — uploads return immediately; the embedding pipeline runs in a daemon thread.
Outbound webhooks — customers register URLs + signing secrets, pick the events they care about (file.uploaded, ask.answered, etc.), and we POST signed JSON. Deliveries are logged with HTTP status + retry count. HMAC-SHA256 in X-Filenergy-Signature: sha256=....
Audit log UI — /audit/ shows every event with type/user/date filters, paginates, and exports CSV. Required for any enterprise sale.
Email service — pluggable adapter (log default, smtp for production). Used for invitations; ready for password resets and notification flows.
Plan-based quotas — free / pro / team tiers with limits on storage, file count, members, and questions per month.
Stripe billing — Checkout sessions, webhook reconciliation, and per-workspace customer/subscription state.
API keys + /api/v1 — bearer-token auth to POST /api/v1/files, POST /api/v1/ask, GET /api/v1/files.
Public share links — per-file unguessable URLs with optional TTL and download caps.
Sliding-window rate limit on /ask (per user) layered on top of the same Event log.
Event log — every meaningful action lands in a row. Substrate for rate-limiter, billing usage, audit log, and webhooks.
Health endpoints — /healthz (cheap liveness) and /readyz (DB + config check) for k8s and uptime monitors.
Settings UI — profile, security, workspace, API keys, webhooks, audit log, billing.
Indexing badges — the file list shows indexed / pending / error per row with one-click reindex.

Security & accounts

2FA (TOTP) — pyotp + QR code. 8 single-use recovery codes generated on enable. Login flow defers login_user until the OTP succeeds, so a stolen password alone is not enough.
WebAuthn / passkeys — full FIDO2 ceremony backed by py_webauthn: navigator.credentials.create/get from the browser, challenge stash in Flask session, attestation + assertion verification server-side. Also ships a stub fallback for tests / pre-JS deploys. Multiple keys per user; per-key labels and last-used timestamps in security settings.
Workspace-wide 2FA enforcement — owners can flip "Require 2FA"; any member without TOTP or a passkey is bounced to /settings/security on the next request until they enroll.
Google SSO — Authlib OpenID Connect. Auto-creates the user (with default workspace) on first login; links to existing accounts by email. Hidden when GOOGLE_OAUTH_CLIENT_ID is unset.
Self-serve account deletion — wipes owned workspaces, files, conversations, API keys, and memberships. Anonymizes events to keep audit trails. Required for GDPR/CCPA compliance.

Bulk operations & exports

Bulk file delete — POST /file/bulk_delete/ with ids[].
Conversation export — GET /ask/c/<id>/export.md returns a Markdown transcript with sources.
Audit log CSV export — GET /audit/export.csv (admin only).
Workspace export (GDPR-grade) — GET /w/export returns a ZIP with every file's bytes, conversation transcripts as Markdown, members CSV, events CSV, and metadata JSON. Owner / admin only.
Personal data export — GET /settings/account/export returns a ZIP with the user's own data + every workspace they own (each as a nested ZIP). Self-serve GDPR portability.
Weekly email digests — Monday recap of last week's uploads, asks, and new members per workspace. Per-user opt-out at /settings/workspace. Wired as flask send-digests for cron / k8s job runners.

Ops

/healthz — liveness, no DB hit.
/readyz — DB SELECT 1 + reports configuration of every external dependency.
/metrics — Prometheus exposition format. Per-endpoint request counters and a duration histogram with the standard SRE buckets.
Structured request logs — every response emits an INFO log with endpoint, status, duration_ms, user_id, workspace_id. Pipe through any log shipper to get user-facing analytics for free.
404 / 500 error pages so the brand stays intact when something breaks.

Self-documenting API

OpenAPI 3 spec at /api/v1/openapi.json.
Swagger UI rendered at /api/v1/docs via the public CDN.

Deploy

Dockerfile with a multi-stage build: a Node 20 stage compiles Tailwind into static/css/app.css, the Python 3.11 stage copies it into the runtime image. base.html links the bundle when present and falls back to the Play CDN in dev. CSP tightens automatically (drops 'unsafe-eval' and the CDN host) when the bundle is in use.
docker-compose.yml with a persistent SQLite volume.
Background jobs with retries — webhook deliveries (5xx + network errors), digest emails (SMTP failures). Exponential backoff (2/4/8/16s) via jobs.enqueue(..., retries=4). RQ-backed when Redis is configured; threading fallback otherwise.

Ingestion

URL ingestion — paste a URL on the upload page; we fetch, strip HTML, extract title/text, and index it like any other file. Caps page size at 10 MB and rejects non-text content types.
OCR fallback — when local extractors return nothing (scanned PDFs, screenshots), we send the file to Claude as a document or image block and ask it to transcribe verbatim. Image MIME types (png/jpg/gif/webp) are now first-class indexable as a result.

First-run experience

Onboarding wizard — new users land on /onboarding/ after registration: name your workspace, optionally seed three sample files so the demo is non-empty.

Activity dashboard

/dashboard/ (owner/admin only) — stat cards (files, indexed, collections, conversations, messages, members, API keys, plan), a 30-day uploads timeseries, a 30-day asks timeseries, and the top-cited files. Plain inline SVG bars — no chart libraries.

Migrations

Alembic via Flask-Migrate — flask db upgrade applies pending migrations in production. Local dev / tests still use db.create_all() so the suite is self-contained. Set FILENERGY_SKIP_CREATE_ALL=1 when running flask db migrate against an empty schema.

Conversation export

/ask/c/<id>/export.md — Markdown
/ask/c/<id>/export.pdf — PDF (via fpdf2, no Cairo dependency)
/ask/c/<id>/export.docx — DOCX (via python-docx, already a dep)

Connectors framework

Pluggable third-party sources via filenergy/services/connectors.py. Manage connections at /connectors/. Each connector implements authorize_url / complete_oauth / sync and runs on demand from the UI or via the scheduler.

Google Drive — full OAuth dance, refresh tokens, native Google docs exported as text/CSV; PDFs and text/* MIME types pulled directly.
Notion — OAuth + search + recursive block flattening into Markdown. Stores each page as <title>.md.
Dropbox — OAuth (offline access), token refresh, list_folder + download for indexable extensions (PDF, DOCX, MD, TXT, CSV, JSON, HTML, log).
Slack — OAuth read-only scopes, fetches the most recent channels and writes per-channel transcripts as slack-<channel>.txt.

Connector sync scheduler

filenergy/services/connector_scheduler.py runs a single daemon thread that wakes every FILENERGY_SYNC_INTERVAL_MIN minutes (default 60), walks every ConnectorAccount whose last_synced_at is stale, and re-runs its sync() as the workspace owner. Errors land in account.last_error instead of crashing the loop. Disabled in TESTING and when FILENERGY_SYNC_SCHEDULER=false.

Browser extension

browser-extension/ — Chrome MV3 popup. Paste your server URL + an API key, click "Save page". POSTs the current tab's URL to /file/from_url/ with Authorization: Bearer …. The endpoint accepts either session cookies or API keys.

Background job queue

filenergy/services/jobs.py — enqueue(target_path, *args).
Default backend is thread: a daemon thread per job (zero-config).
Set FILENERGY_JOBS_BACKEND=rq + REDIS_URL=redis://... to push to a Redis queue (run an rq worker against it). Falls back to thread when the dep or URL is missing.
Tests force synchronous execution via app.config["TESTING"].

Postgres + pgvector

DB engine is whatever URI you put in FILENERGY_DB_URI. Set postgresql://user:pass@host/db, then flask db upgrade.
For real-scale RAG retrieval, install pgvector>=0.2 and run:
```
from filenergy.services import pgvector_store
pgvector_store.enable_pgvector(dim=512)   # creates extension + col + ivfflat index
pgvector_store.reembed_existing()         # back-fill from JSON column
```
After that, embeddings.search automatically uses an ORDER BY cosine SQL query instead of pulling every chunk into Python. SQLite callers stay on the JSON+numpy path with no code change.

SAML SSO

filenergy/services/saml_sso.py is a real python3-saml wrapper (not a stub). Set SAML_ENABLED=true, SAML_IDP_METADATA_URL, SAML_SP_ENTITY_ID, and (optionally) SAML_SP_X509_CERT / SAML_SP_PRIVATE_KEY. The Dockerfile installs the required libxml2 / xmlsec1 system libraries.
/saml/login redirects the browser to the IdP. /saml/acs validates the SAMLResponse, provisions the user (or links by email), creates a default workspace, and logs them in.
/saml/status reports configuration + whether python3-saml is importable, so you can verify env wiring before flipping it on.

Browser extension

browser-extension/ — Chrome MV3 popup. Paste your server URL + an API key, click "Save page". POSTs the current tab's URL to /file/from_url/ with Authorization: Bearer …. The endpoint accepts either session cookies or API keys.

Encryption at rest

services/crypto.py exposes a Fernet-backed EncryptedText SQLAlchemy type. Sensitive columns — File.text_content, Chunk.embedding, ConnectorAccount.access_token/refresh_token, User.totp_secret — go through it.
Activate by setting FILENERGY_ENCRYPTION_KEY (mint one with python manage.py generate-encryption-key). Without the key, columns round-trip as plaintext — backwards compatible with old data and dev environments.
New writes get an enc: prefix; old plaintext rows coexist with new encrypted ones. Run python manage.py reencrypt once after enabling to back-fill.

Chunk-level provenance

Every assistant turn now creates MessageCitation rows linking the Message to each retrieved Chunk + the cosine score. The dashboard surfaces a "Most-cited chunks" panel with the snippet inline, and the "Most-cited files" panel now uses real citation counts instead of the prior coarse ask.answered event proxy.

Incremental connector sync

ConnectorAccount.sync_cursor is the resume token; each connector uses its native cursor primitive:

Drive — q=modifiedTime > '<RFC3339>', cursor = newest modifiedTime from the latest page.
Notion — start_cursor / next_cursor. Cleared when the result set is exhausted so the next tick starts over.
Dropbox — first call hits /2/files/list_folder, subsequent calls hit /list_folder/continue with the saved cursor (delta-native).
Slack — oldest=<ts>. New deltas are appended to the existing per-channel transcript file rather than written as fresh files.

Security hardening

CSRF protection on every browser POST via Flask-WTF. The /api/v1, /webhooks/stripe, and /saml/acs blueprints opt out because they authenticate with Bearer / HMAC / SAML signature, not session cookies. JSON Ajax requests pick up the token from a <meta> tag and send it as X-CSRFToken.
Failed-login rate limit keyed on the email being attempted. Configurable via FILENERGY_LOGIN_RATE_LIMIT / FILENERGY_LOGIN_RATE_WINDOW. Bucket per email so one attacker can't lock out a real user.
Security headers on every response: Strict-Transport-Security, X-Content-Type-Options, X-Frame-Options: DENY, Referrer-Policy, Permissions-Policy, and a Content-Security-Policy that allows Bootstrap-3 inline scripts + the Swagger UI CDN. Set FILENERGY_DISABLE_HSTS=true for plain-HTTP dev deploys.
Session management via a UserSession row per active browser. /settings/security lists them with browser / IP / last-seen, lets the user revoke any one, or "Log out of all other sessions". The middleware refuses cookies whose row was revoked.

Conversation share links

POST /ask/c/<id>/share (or POST /api/v1/conversations/<id>/share-links) mints a public read-only URL at /sc/<token>.
TTL via ttl_hours; revocable; view_count increments on every successful landing render.

Source paragraph viewer

Indexer now stores char_offset_start / char_offset_end on every Chunk so we can recover the exact span inside File.text_content.
/file/chunk/<id>/context returns the cited span + ~280 chars of surrounding context. The dashboard's "Most-cited chunks" panel uses it to expand a citation inline with the span highlighted.

Per-API-key scopes

ApiKey.scopes (e.g. ["files:read", "ask:write"]) restricts which endpoints a key can hit. Empty = full access (back-compat).
The /settings/keys mint form has checkboxes for every recognised scope. files:write implies files:read.

`/api/v1` is now a real API

In addition to /files and /ask:

GET/POST/DELETE /api/v1/files/<id>
GET/POST/DELETE /api/v1/collections, PUT /api/v1/collections/<id>/files/<id>
GET /api/v1/conversations, GET/DELETE /api/v1/conversations/<id>
POST /api/v1/files/<id>/share-links, DELETE /api/v1/share-links/<id>
POST /api/v1/conversations/<id>/share-links
GET/POST/DELETE /api/v1/webhooks
GET /api/v1/members, POST /api/v1/invitations

Every endpoint enforces the relevant scope.

Tests

718 tests, 97.3% coverage (pytest + pytest-cov).

Where we sit vs the competition

Capability	Filenergy	NotebookLM	Humata	Mendable	Glean
Self-host (Docker)	✅	❌	❌	❌	❌
Multi-tenant workspaces	✅	❌	❌	✅	✅
Collections / per-doc chat	✅	✅	✅	❌	❌
Streaming RAG with citations	✅	✅	✅	✅	✅
Auto-summary + suggested Qs	✅	✅	✅	❌	❌
OCR for scanned PDFs / images	✅	✅	✅	❌	❌
URL ingestion	✅	✅	❌	❌	❌
Conversation export (MD/PDF/DOCX)	✅	partial	❌	❌	❌
GDrive / Notion / Dropbox / Slack connectors	✅	❌	❌	✅	✅
Connector sync scheduler	✅	❌	❌	✅	✅
Browser extension	✅	❌	❌	❌	❌
Background job queue (RQ)	✅	n/a	n/a	n/a	n/a
pgvector retrieval	✅	n/a	n/a	n/a	n/a
Outbound webhooks	✅	❌	❌	✅	✅
Audit log UI + CSV export	✅	❌	❌	❌	✅
Activity dashboard	✅	❌	❌	✅	✅
Stripe-billed plan tiers	✅	n/a	✅	✅	✅
Public REST API + OpenAPI	✅	❌	✅	✅	✅
Public share links with TTL	✅	❌	❌	❌	❌
2FA (TOTP)	✅	❌	❌	❌	✅
SSO (Google OIDC)	✅	✅	✅	✅	✅
Self-serve account deletion	✅	✅	❌	❌	❌
Prometheus `/metrics`	✅	❌	❌	❌	✅
Onboarding wizard	✅	✅	✅	❌	✅
Alembic migrations	✅	n/a	n/a	n/a	n/a
SAML SSO	✅	❌	❌	✅	✅

Setup

1. Install

pip install -r requirements.txt

2. Configure (`.env`)

# Required for /ask
ANTHROPIC_API_KEY=sk-ant-...
# Required for indexing + retrieval
VOYAGE_API_KEY=pa-...

# Stripe (optional — billing UI degrades gracefully without it)
STRIPE_SECRET_KEY=sk_test_...
STRIPE_PUBLIC_KEY=pk_test_...
STRIPE_WEBHOOK_SECRET=whsec_...
STRIPE_PRICE_PRO=price_...
STRIPE_PRICE_TEAM=price_...
FILENERGY_BASE_URL=https://your-domain.com

# Google OAuth (optional — "Sign in with Google" hides without these)
GOOGLE_OAUTH_CLIENT_ID=...
GOOGLE_OAUTH_CLIENT_SECRET=...

# Email (optional — log adapter prints to stderr; smtp sends real mail)
FILENERGY_EMAIL_ADAPTER=log     # or smtp
FILENERGY_EMAIL_FROM=filenergy@yourdomain.com
FILENERGY_SMTP_HOST=smtp.example.com
FILENERGY_SMTP_PORT=587
FILENERGY_SMTP_USER=...
FILENERGY_SMTP_PASSWORD=...
FILENERGY_SMTP_TLS=true

# Misc
FILENERGY_SECRET_KEY=change-me
FILENERGY_DB_PATH=filenergy.db
FILENERGY_UPLOAD_DIR=files
CLAUDE_MODEL=claude-opus-4-7
VOYAGE_EMBED_MODEL=voyage-3-lite
FILENERGY_SYNC_INDEXING=false        # true = inline indexing (helpful in tests)
FILENERGY_ASK_RATE_LIMIT=30          # per window
FILENERGY_ASK_RATE_WINDOW=60         # seconds

3. Run

python manage.py

http://localhost:5000

For the Stripe webhook locally:

stripe listen --forward-to localhost:5000/webhooks/stripe

CLI

python manage.py create-superuser admin@example.com 'a-good-password'
python manage.py reindex

Plans

	Free	Pro $19/mo	Team $99/mo
Questions / month	100	2,000	20,000
Storage	100 MB	5 GB	100 GB
Files	25	1,000	25,000
Members	1	1	25

Adjust in filenergy/settings.py:PLAN_LIMITS.

API

# Mint a key in /settings/keys, then:

curl -X POST http://localhost:5000/api/v1/ask \
  -H "Authorization: Bearer $FILENERGY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"question": "What do my contracts say about termination?"}'

curl -X POST http://localhost:5000/api/v1/files \
  -H "Authorization: Bearer $FILENERGY_TOKEN" \
  -F "files[]=@./report.pdf"

curl -H "Authorization: Bearer $FILENERGY_TOKEN" \
  http://localhost:5000/api/v1/files

Errors:

Code	Meaning
401	Invalid / missing API key
402	Plan quota exceeded (file count, storage, asks/month)
429	Rate limit hit (per-user sliding window on `/ask`)
503	Anthropic / Voyage / Stripe not configured on server

Architecture

upload  →  text extraction  →  chunk (1200/150)  →  Voyage embed
                                                       │
                                                       ▼
                                                  SQLite

ask     →  Voyage embed  →  cosine top-K (workspace scope)  →  Claude (RAG, streaming)
                                                                  │
                                                                  ▼
                                                              SSE → browser
                                                              persisted to thread

billing →  webhook  →  workspace.plan flip  →  next quota check sees the new plan

Tenant boundary is workspace_id on File, Conversation, Event, ApiKey. All queries filter on it.
Embeddings are JSON-encoded float arrays in SQLite. Fine up to ~10K chunks per workspace; swap in sqlite-vec or pgvector for more.
Background indexing runs in a daemon thread with its own app context. Tests force it inline via app.config["TESTING"].
Rate limiter is a sliding window over the Event table — same source of truth as billing usage gauges.

Project layout

filenergy/
├── settings.py                 # env config + PLAN_LIMITS
├── __init__.py                 # app, db, login_manager
├── middleware.py               # g.user / g.workspace
├── admin.py                    # superuser-only Flask-Admin
├── models/
│   └── __init__.py             # User, Workspace, WorkspaceMember,
│                               # WorkspaceInvitation, File, Chunk,
│                               # Conversation, Message, Event, ApiKey,
│                               # ShareLink
├── services/
│   ├── base.py                 # generic SQLAlchemy service
│   ├── user.py                 # auth + register/login (auto-creates ws)
│   ├── workspaces.py           # tenancy, invitations, switching
│   ├── api_keys.py             # token mint/verify/revoke
│   ├── share_links.py          # public share TTL + cap
│   ├── billing.py              # Stripe + plan-quota checks
│   ├── file.py                 # upload, async index, search
│   ├── extraction.py           # pdf/docx/txt extractors + chunker
│   ├── embeddings.py           # Voyage client + cosine retrieval
│   ├── chat.py                 # RAG + Claude streaming
│   ├── conversations.py        # threads + messages
│   ├── events.py               # analytics + audit
│   └── rate_limit.py           # DB-backed sliding window
├── views/
│   ├── index.py
│   ├── user.py
│   ├── file.py                 # CRUD, reindex, share
│   ├── ask.py                  # JSON ask + SSE stream
│   ├── workspace.py            # switch, invite, accept, members
│   ├── settings_views.py       # profile, keys, ws, billing
│   ├── share.py                # /s/<token>
│   ├── api_v1.py               # bearer-auth API
│   └── billing.py              # /webhooks/stripe
├── templates/                  # bootstrap-3 templates
└── static/

tests/                          # pytest suite, 272 tests, 97.6% coverage

Testing

pip install pytest pytest-cov stripe
python -m pytest                  # run the suite
python -m pytest --cov            # with coverage report

Stubs in tests/conftest.py replace Voyage and Anthropic clients; Stripe is faked per-test via sys.modules. Zero network.

Deploy

docker compose up --build

Then point a reverse proxy (nginx, Caddy, Cloudflare) at port 5000 with TLS terminated.

Roadmap

Per-connector OAuth scope tightening (Slack still asks groups:* scopes; Notion + Dropbox could be narrower).
KEK rotation via MultiFernet so secrets can rotate without a full re-encrypt pass.
Citation drill-through — click a chunk on the dashboard and jump to the source paragraph in the file detail view.
Cron / RQ-driven scheduler (today's daemon thread is fine for one Filenergy process; scale-out needs a real worker).

Name		Name	Last commit message	Last commit date
Latest commit History 43 Commits
browser-extension		browser-extension
docs/screenshots		docs/screenshots
filenergy		filenergy
migrations		migrations
screenshots		screenshots
tests		tests
.coveragerc		.coveragerc
.dockerignore		.dockerignore
.gitignore		.gitignore
Dockerfile		Dockerfile
README.md		README.md
docker-compose.yml		docker-compose.yml
manage.py		manage.py
package.json		package.json
pytest.ini		pytest.ini
requirements.txt		requirements.txt
tailwind.config.js		tailwind.config.js

Folders and files

Latest commit

History

Repository files navigation

Filenergy

Screenshots

What's new in this release

UX

Highlights

Security & accounts

Bulk operations & exports

Ops

Self-documenting API

Deploy

Ingestion

First-run experience

Activity dashboard

Migrations

Conversation export

Connectors framework

Connector sync scheduler

Browser extension

Background job queue

Postgres + pgvector

SAML SSO

Browser extension

Encryption at rest

Chunk-level provenance

Incremental connector sync

Security hardening

Conversation share links

Source paragraph viewer

Per-API-key scopes

/api/v1 is now a real API

Tests

Where we sit vs the competition

Setup

1. Install

2. Configure (.env)

3. Run

CLI

Plans

API

Architecture

Project layout

Testing

Deploy

Roadmap

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Uh oh!

Contributors

Uh oh!

Languages

`/api/v1` is now a real API

2. Configure (`.env`)

Packages