LeadGrow · Video Pipeline

tella-cli

Command-line control over your Tella library. 306 videos. Agent-native. Offline FTS5 search. No more clicking through thumbnails.

Python 3.11 SQLite + FTS5 Auto-JSON when piped Typed exit codes ~/.tella/db.sqlite CLI Printing Press

01 / 10

The Problem

306 videos. No way to query them.

Tella's web app is the only interface. Finding last week's recording, auditing view counts, or piping video IDs into a pipeline meant scrolling through pages of thumbnails — manually.

306

Videos in Tella

948

Total Views

3.1

Avg Views / Video

→ all queryable in <2 seconds, offline

02 / 10

Architecture

Two layers. One sync.

🎬

Tella API

api.tella.com/v1
Bearer token
100 req/min

→

⚡

tella sync

Full paginated pull
Auto-loads .env
~2 sec / 300 videos

→

🗄

SQLite + FTS5

~/.tella/db.sqlite
Offline instant
Raw SQL access

# one-time setup TELLA_API_KEY=tella_pk_... # in .env or shell tella sync # 306 videos → SQLite in ~2 sec # everything below is instant and offline tella search "product demo" tella videos top -n 10 tella sql "SELECT name, views FROM videos ORDER BY views DESC"

03 / 10

Commands

Full command reference

Command	Source	Description
videos list	Live API	Paginated list. `--all` exhausts pages, `--playlist ID` filters
videos recent -n N	SQLite	Last N by created_at. Sync first.
videos top -n N	SQLite	Top N by view count. Content audit.
playlists list	Live API	Personal or org playlists (`--visibility org`)
sync	Live → SQLite	Full pull of all videos + playlists into local cache
search <query>	SQLite FTS5	Offline FTS. AND / OR / NOT / "phrases" / prefix*
stats	SQLite	Total videos, views, avg views, last sync time
health	Live API	Auth + rate limit headroom check
sql "<SELECT>"	SQLite	Raw read-only SQL escape hatch. SELECT / WITH only.

04 / 10

CLI Printing Press Design

Built for agents. Loved by humans.

Every flag and exit code exists because AI agents invoke this thousands of times daily.

Universal Flags

--json Force JSON output

Auto when stdout piped. No flag needed.

--compact Key fields only

~60% fewer tokens. id, name, views, created_at, view_page.

--quiet Suppress stderr

Clean pipelines. Errors to stderr, data to stdout.

Typed Exit Codes

0

success

Operation completed

2

usage

Bad flag or invalid SQL

3

not_found

Empty results or empty cache

4

auth

Missing or invalid API key

5

api_error

4xx / 5xx from Tella

7

rate_limited

429 — back off and retry

05 / 10

Use Cases

9 workflows for LeadGrow video

01 — Morning Kickoff

Session start

Sync fresh state, see what's waiting to edit.

tella sync --quiet
tella videos recent -n 5

02 — Search

Find by topic

FTS5 across all 306 video names + descriptions.

tella search "product demo"
tella search "gtm OR outbound"

03 — Unedited Audit

This week's recordings

Find recordings from last 7 days.

tella sql "SELECT name FROM videos WHERE created_at >= date('now','-7 days')"

04 — Performance

View count audit

Which videos are getting traction?

tella videos top -n 10
tella stats

05 — Playlist Scoped

Filter by playlist

All videos in a Tella playlist.

tella playlists list
tella videos list --playlist <id> --all

06 — Agent Pipeline

Full JSON pipeline

Pipe into scripts or Claude Code sessions.

tella videos recent -n 20 --json --compact | python process.py

07 — Pre-Batch Check

Health before ops

Verify auth + rate limits before bulk work.

tella health --json
# exit 0 = proceed

08 — Re-Share

Zero-view videos

Published but never seen. Candidates to resurface.

tella sql "SELECT name FROM videos WHERE views = 0"

09 — Weekly Rhythm

Output by week

Recording cadence for the month.

tella sql "SELECT strftime('%W', created_at) w, COUNT(*) n FROM videos GROUP BY w"

06 / 10

Offline Power

FTS5 search + raw SQL

Search syntax

Pattern	Meaning
product demo	Both words, any order
"product demo"	Exact phrase
gtm OR outbound	Either word
recording NOT personal	Exclude term
outbound*	Prefix match

→ ranked by relevance, instant offline

Power SQL examples

-- top performers this year SELECT name, views FROM videos WHERE created_at >= '2026-01-01' ORDER BY views DESC LIMIT 10; -- recording velocity by month SELECT strftime('%Y-%m', created_at) mo, COUNT(*) n FROM videos GROUP BY mo ORDER BY mo DESC; -- zero-engagement, newest first SELECT id, name, created_at FROM videos WHERE views = 0 ORDER BY created_at DESC;

07 / 10

Live Data — Synced 2026-05-11

What's in the vault

306

Total Videos

948

Total Views

3.1

Avg Views

54

CLI Printing Press principles

The framework that climbs the five-rung creativity ladder from endpoint-wrapper to behavioral insight tool.

Absorption + Transcendence

Catalogs competitors, absorbs their features, then adds novel capabilities (offline FTS5, behavioral analytics) that stateless wrappers can't match.

Creativity Ladder — Rung 5

Most CLIs stop at Rung 1 (wrap the endpoint). tella-cli reaches Rung 5: videos top, weekly cadence SQL, zero-view audit.

Agent-Native by Default

Auto-JSON when piped. --compact cuts token spend 60%. Six typed exit codes replace brittle stderr parsing. Every design decision serves the agent first.

Two-Tier Data Layer

Live API for fresh state. SQLite + FTS5 for instant offline queries. Sync once, query forever — no rate limit pressure on your analytical loops.

Framework: C:\Users\mitch\Everything_CC\cli-printing-press\ · Tella catalog: catalog/tella.yaml

09 / 10

Get Started

Up in 60 seconds

# 1. add to leadgrow-video/.env TELLA_API_KEY=tella_pk_your_key_here # 2. sync 306 videos to local SQLite npm run tella -- sync # 3. explore npm run tella -- videos recent -n 10 npm run tella -- search "product demo" npm run tella -- videos top -n 5 npm run tella -- stats npm run tella -- health # or direct (auto-loads .env from cwd) python C:/Users/mitch/Everything_CC/cli/tella-cli/tella.py sync

306 Videos FTS5 Offline CLI Printing Press Agent-Native

Pipeline is oxygen. Speed is the moat.

10 / 10