Back topod. pod docs

Command reference

Every command pod understands, organized by what you're trying to do. You never actually need to memorize any of them, though, because plain English always works.

Note: Once you type pod and you're inside the studio (the bordered box with the ❯), you never need the pod prefix: just type resume, not pod resume. And you can always skip commands entirely: say "make a 30-second teaser for my thriller, telugu, moody" and the front desk turns your words into the exact command, shows it to you, and waits. Enter runs it, e lets you edit it, n means never mind. Talking is the intended way; commands are here for when you want precision. More on this in Talking to pod.
Tip: Tab completes command words and your episode names, so you rarely have to type a full name. Up/down arrows browse your history only when the line is empty. They can never delete what you've typed.

Start something

pod

Opens the studio. This is the one command you type in your Terminal (the app where you type text and press Enter to run it); everything after that happens inside pod's own friendly input box.

❯ pod
  (the studio opens — talk, or type any command below)

pod ui

Opens Pod Studio (pod's browser view) on its own, showing your production as visual cards with a chat box that's the same front desk. Running plain pod already opens it automatically alongside the Terminal; use pod ui when you want just the page. It's local and private (it binds to 127.0.0.1, your machine only, and nothing is served to the network). More in talking to pod.

❯ pod ui
  (Pod Studio opens in your browser — the canvas + chat)

To keep pod terminal-only and never open the page, start it as POD_NO_UI=1 pod (both words on the same line, Mac/Linux).

pod "an idea"

You can also hand pod an idea straight from the Terminal. The front desk turns your plain language into the right command and shows it before running anything.

❯ pod "two robots fall in love at a bus stop"

produce "<brief or file>"

The main event: give pod an idea, a pasted script, or a screenplay file, and it writes, casts, designs, shoots, quality-checks and edits a finished film. It takes flags for everything you'd otherwise be asked in the upfront brief:

❯ produce "Show V4 EP 1-30.pdf" --episode 2 --series basti --duration 90 --aspect 9:16 --language telugu

You don't need most of these. pod asks for anything missing, once, upfront, and remembers "your usual" after a couple of productions.

explore "<subject>"

For when you don't have the idea yet. Pod pitches 4–5 distinct concepts with looks and formats; pick one and it goes straight to production.

❯ explore "brainstorm some ideas for a coffee brand ad"
  (4-5 concepts appear — pick one, pod produces it)

cast "idea"

A casting-and-price preview only. pod's casting director picks generation engines for the idea and shows what it would cost, without producing anything (about 1 credit).

❯ cast "moody telugu vertical teaser, 60 seconds"

presets

Shows pod's five proven formats: ready-made shapes for common jobs, each carrying its own length, aspect and delivery quality. Start a film on one with --preset:

❯ produce "coffee subscription, launch week" --preset ugc

A trailer isn't a preset. It's --teaser, which shortens the film and protects the ending.

demo

Renders a free 5-second clip entirely on your machine, with no credits and no internet generation, to prove the setup works. Great right after install.

Continue something

episodes

Shows what's on the lot: every film you've started or finished.

status <name>

Shows a film's progress and (the useful part) exactly what a hand-edit has made stale. Edit the brief, the script, a character sheet, or a single shot's prompt (by hand, or with the ✎ editors in Pod Studio) and status tells you which shots now need re-rendering, with a rough price and the exact redo command to apply it. It's read-only: it never regenerates anything and never spends a credit. Pod Studio shows the same thing visually: the affected cards get an amber "stale" badge, while everything still valid stays green, so you can see precisely what you'd keep. Nothing re-renders until you ask and approve the price.

❯ status basti-ep2
  edited since the last run — now stale:
    • sheet:char-ravi
  affected shots: shot-02 (ravi changed) — re-render still+clip
  ~430 credits to regenerate — rough; the exact figure shows at the cost gate,
  and nothing runs until you confirm
  apply per shot: pod redo basti-ep2 --shot shot-02

resume [name]

Picks a film back up from its exact checkpoint, after a stop, a crash, a closed laptop, or just tomorrow morning. Approved work is never redone, and paid work is never paid for twice. It resumes in full control, right where you left off:

❯ resume basti-ep2 --steer
  (continues mid-film, with you back at the wheel)

finalize <name>

Masters a drafts-only film at full resolution. This is the second half of the "d" choice at the cost gate: watch a cheap 480p cut first, then finalize when you're happy.

❯ finalize basti-ep2

open [name]

Plays a finished film in your usual video player. With no name it opens the newest film you've finished, the fastest way to watch what you just made.

❯ open

Fix something

edit <name> "note"

Re-cuts the film from existing takes with a plain-language edit note. Edits are free, because nothing re-renders.

❯ edit basti-ep2 "cut the first 3 seconds, pacier"

redo <name>

Re-shoots a shot, or re-specs the film. Unlike edits, re-shoots cost credits, so pod always shows the price and asks first.

❯ redo basti-ep2 --shot shot-07

--from <stage> rewinds the production to an earlier point and runs forward from there. The stages, in pipeline order, are script, casting, characters, shots, prompts, cost-gate, shoot, edit and report. If you think in terms of stills or clips, both mean shoot.

--duration and --language apply to the whole film, never one shot: a new length re-writes the script and re-cuts the shots, and a new language re-shoots everything. If you only want the voices in another language on a film that's already finished, that's localize, and it's far cheaper.

fixclip <name> --shot …

Targeted surgery on one clip when only a piece of it needs attention.

takes <name>

Browses the alternative takes pod kept for each shot. Sometimes the fix is a take you already have. Add --shot shot-NN --use take-N to make one of them the active take: on a finished film pod re-stitches the cut from footage you already own, so switching takes is free, with no new generation. In Pod Studio the same thing is a click: 🎞 takes on a clip card, then use this take.

❯ takes basti-teaser --shot shot-01 --use take-2
shot-01 now uses take-2 · re-stitching from existing footage (free)… done.

Audio & language

dub <name> --shot shot-03 --audio my.mp3

Swaps one shot's audio for a real recording (your actor's voice, or your own) and re-stitches the film. The house workflow: ship first, watch, then dub the lines that deserve a human.

vo <name>

Regenerates the voice-over for a film.

localize <name> --language hindi

Makes a dubbed version of a finished film in another language: one telugu vertical becomes a hindi one without re-shooting a frame.

Money

credits

Your generation balance, right now.

estimate --runtime 120

A cost preview for a film of that length, before you commit to anything.

❯ estimate --runtime 90
  (price options appear — nothing is spent)

Remember: the cost gate always shows the exact estimate before any footage renders, drafts cost roughly a fifth of masters, and a bare "no" never spends money. Full picture in Money & credits.

Setup & housekeeping

doctor

Checks and fixes the whole setup, and renders two free local test clips. Run it whenever anything looks off: wrong-looking titles, missing tools, general weirdness.

login

Re-runs the friendly key wizard for your license key, generation key (kie.ai/api-key for own-key users; managed accounts enter only a license key), and the optional ElevenLabs voice key. See Setting up your keys.

update

Fetches the latest pod. Every update is cryptographically signed: pod verifies the release came from its makers and refuses anything tampered, mismatched, or older than what you have. On Windows it finishes with one move command shown for you to run.

series [name] / library [name]

series opens your show's "bible", the recurring faces, voices and look that carry from episode 1 to episode 30. library lists the films you've produced inside a series. Both are read-only and free. See Series & continuity.

❯ series basti
❯ library basti

templates

The motion-graphics menu: the animated titles, subtitles and end cards pod can dress a finished film with. Graphics render locally on your own machine, so they cost nothing.

version

Prints the version of pod you're running, handy when you're reporting a bug.

models

Advanced, and honestly optional: shows which AI does the thinking for each crew seat, and lets you reseat them (models writer=opus; aliases include opus, fable, sonnet, gpt-5.5, flash, gemini-pro; models reset restores defaults). Picks persist across restarts and updates, and pod refuses unsafe picks for seats that need to see images. The defaults are tuned, so change them only if you know why.

feedback / report

feedback sends a thought to pod's makers; report sends a bug report with context attached. Fixes ship frequently, often the same week.

help / exit

help any time you're lost; exit leaves the studio. And Ctrl-C cancels a running job without kicking you out. Progress is checkpointed, so nothing is lost.

Not on this page: the commands that file an episode under a client project and sync it to a studio hub: pod projects, pod push, pod pull, pod notes. They're real and they work, but only inside a studio whose owner has set the hub up on your machine; on any other copy of pod they answer "hub not configured" and stop. They're documented where they make sense, on hub, portal & the terminal.

New here? Start with the quickstart, then make your first film. When you're reviewing shots, the camera language page teaches you the vocabulary the crew speaks.