Aide/.env.example at main · 0dillon/Aide · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
# Copy this file to `.env` and fill from your Monnify dashboard → Developer page.
# NEVER commit the real .env.

# Developer → API Keys & Contracts → "Show Keys"
MONNIFY_API_KEY=MK_TEST_xxxxxxxxxx
MONNIFY_SECRET_KEY=xxxxxxxxxxxxxxxxxxxx

# Sandbox base URL (Developer page shows it; sandbox is usually this)
MONNIFY_BASE_URL=https://sandbox.monnify.com

# Developer → Contract Code
MONNIFY_CONTRACT_CODE=4523534626

# Developer → Wallet Account Number (source account for disbursements)
MONNIFY_WALLET_ACCOUNT_NUMBER=

# A test destination for the disbursement proof.
# Use a known Monnify sandbox test account, or one of your own reserved accounts.
TEST_DEST_ACCOUNT_NUMBER=0000000000
TEST_DEST_BANK_CODE=058

# Port for the local webhook receiver
WEBHOOK_PORT=4000

# ---- Convex (all of Aide's data) ----
# You do NOT set these by hand. Run `npx convex dev` in a second terminal and it
# creates a deployment and writes CONVEX_DEPLOYMENT + NEXT_PUBLIC_CONVEX_URL
# into .env.local for you. Leave that terminal running while you develop — it
# also pushes changes in convex/ to the backend.
# The demo worker and employer accounts seed themselves on first request, so a
# brand-new deployment works immediately with no extra step.

# ---- Aide agent (voice loop) ----
# DeepSeek API key for the Aide agent. Get one at platform.deepseek.com.
# DeepSeek is OpenAI-compatible, supports tool calling, and is far cheaper than Claude.
DEEPSEEK_API_KEY=sk-xxxxxxxx
# Optional model override (defaults to deepseek-chat; must support tool calling)
AIDE_MODEL=deepseek-chat

# ---- Neural text-to-speech (Edge TTS via Python subprocess) ----
# Requires Python 3 + `pip install edge-tts` on this machine (see
# requirements.txt). No API key needed — this calls Microsoft's free Edge
# voices through Python's edge_tts library, which (unlike the Node ws-based
# npm package) isn't blocked with a 403.
# Optional voice override — en-NG-EzinneNeural (female) / en-NG-AbeoNeural
# (male) are the only two Nigerian English neural voices in the catalog.
EDGE_TTS_VOICE=en-NG-EzinneNeural
# Optional override if `python` isn't on PATH (e.g. "python3", or a full path).
PYTHON_BIN=python

# Where the browser fetches speech from. Leave UNSET locally: the Next route at
# /api/tts keeps one warm Python subprocess alive, which is much faster for
# development. In production it is set to /api/speak, the native Python
# serverless function — a serverless function can't hold a child process.
# NEXT_PUBLIC_TTS_PATH=/api/speak