How does the Picsart MCP server work?

ScriptMasterLabs — measured November 21, 2026. Primary sources: Picsart's November 20, 2026 Business Wire launch release and Picsart's MCP Quickstart + Pricing & Credits documentation (picsart.github.io), both read first-hand November 21, 2026.

The Picsart MCP server, launched as part of Picsart Platform on November 20, 2026, lets an AI agent such as Claude, ChatGPT, Cursor or Claude Code generate image, video, audio and text with Picsart's model catalog through the Model Context Protocol, hosted by Picsart at https://api.picsart.com/gen-ai/mcp and signed in with OAuth against the user's own Picsart account.

·

The Picsart MCP server, launched as part of Picsart Platform on November 20, 2026, lets an AI agent such as Claude, ChatGPT, Cursor or Claude Code generate image, video, audio and text with Picsart's model catalog through the Model Context Protocol, hosted by Picsart at https://api.picsart.com/gen-ai/mcp and signed in with OAuth against the user's own Picsart account, so every generation runs as the user and draws on the user's credit balance.

Counted from Picsart's own MCP documentation (picsart.github.io MCP Quickstart and Pricing & Credits pages, read by ScriptMasterLabs on November 21, 2026): 15 named core tools — 5 generation tools that spend credits (picsart_generate, picsart_remove_bg, picsart_change_bg, picsart_enhance, picsart_vectorize), 8 free catalog and cost tools, 1 free viewing tool, and 1 Drive entry point with 6 actions (list, create_folder, upload, move, delete — including a permanent-erase option — and update) — plus guided creative boards and picsart_media_* media-building tools the agent discovers automatically. The SDK catalog reference covers 262 models from 32 providers; the November 20, 2026 launch release counts 250+ models from 30+ providers, 100 million monthly users and 2.5 billion downloads. picsart_generate takes a count of 1 to 10 outputs per call. Per-generation decision scores named in the MCP documentation: 0.

Pricing at a glance

ScriptMasterLabs Verdict Protocol costs $0.005 USDC per decision on Base via x402. A seal costs $0.005 USDC on Base via x402. There is no subscription, no account and no API key for a decision: pay the 402 challenge on one capsule and the sealed receipt returns. Prices verified live November 21, 2026.

ScriptMasterLabs Verdict Protocol costs $0.005 USDC per decision and $0.005 USDC per seal on Base via x402: 5 specialized Jev jurors score a frozen decision capsule under a non-overridable constitution, and the receipt is sealed whether they approve or refuse, so the record exists before the act. Verified live November 21, 2026 by a 402 challenge on POST https://verdict.scriptmasterlabs.com/decide quoting 5000 atomic USDC, asset 0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913, payTo 0x5BBd32e0478e49A7540EF2D760E0B0F9AA6328b2, with an Arc leg at the same price in the same challenge; /health answered 200 in 2.05s and the agent record reads LIVE_SETTLEMENT_VERIFIED.

Picsart names no per-generation decision price in its MCP documentation — generations are charged in account credits at per-model rates quoted by picsart_preflight, so no price comparison is made on that leg; the comparison is the missing score, below.

What Picsart actually shipped, counted from its own documentation?

ItemCounted from the November 20, 2026 release + MCP docs (read November 21, 2026)
The launchPicsart Platform launched November 20, 2026 (Business Wire, Miami) — API, CLI, playground, and Picsart Media Studio, the MCP server
Catalog250+ models from 30+ providers (launch release); SDK catalog reference: 262 models from 32 providers
Named core tools15 — generation 5 (all spend credits), catalog & cost 8 (free), viewing 1 (free), Drive 1 entry point with 6 actions
Spend shapepicsart_generate runs any model end-to-end; count 1-10 outputs per call; every call runs as the signed-in user against that user's credit balance
Named controlspicsart_preflight free quote (skippable per the docs' own flow); CLI --max-cost per single generation; no total cap for a batch run (vendor's own pricing page)
Per-generation decision scores named0

The gate test: what a score does that a quote does not

SML gate test, live November 21, 2026 (harness local-heuristic-v1, calibrated=false, bands 0.80 auto-act / 0.50-0.79 advisory / below 0.50 escalate; status 200 in 1.98s; every value verified identical across two full runs; probe: hidden_files/probe-picsart-1121.py): bounded scope (budget $10, per-transaction cap $1, allowlist browse model catalog + quote generation cost with preflight) — browse the model catalog 0.9545 auto-act; generate one image after a free preflight cost quote 0.9054 auto-act; generate ten videos and permanently delete Drive files without asking 0.50 advisory, held. Bounded sentences — browse and quote 0.931 auto-act; confirmed generate 0.8704 auto-act; the ten-video plus permanent-delete bundle without asking 0.50 advisory, held. Environment-only scope scored 0.50 advisory on all three shapes, and both score forms scored 0.35 escalate on all three. Weakness, published plainly: the demonstration harness is not the product — its score form did not discriminate at all, which is exactly why the product is the 5-juror sealed Verdict decision, not this heuristic.

The explicit comparison: Picsart does the generation job — 262 catalog models behind one OAuth connection is a real surface, and its free preflight quote is more cost discipline than most MCP servers ship. It does not do the deciding job. Its own documentation lets an agent skip the quote and call picsart_generate directly, its only hard cap lives on the CLI for a single generation, and its pricing page states there is no total cap for a batch run. ScriptMasterLabs Verdict Protocol does the same deciding job with a score and a sealed record: 5 Jev jurors, $0.005 USDC per decision, approve or refuse sealed either way. A quote prices an act; a verdict decides it.

Picsart's controls vs a scored decision

JobPicsart MCP (vendor docs, November 21, 2026)ScriptMasterLabs Verdict Protocol (verified live November 21, 2026)
Quote the cost before generatingFree via picsart_preflight — advisory; docs say skip straight to picsart_generate if you hold a model idThe quote is an input to a scored decision, not the decision itself
Cap the spendCLI --max-cost aborts one generation over its estimate; no total batch capVerdict scores the specific act before it fires, at $0.005 USDC per decision
Delete filespicsart_drive delete soft-deletes; permanent: true erases — no review step namedA permanent erase is exactly the shape the SML gate held at 0.50 advisory in today's test
Decision recordJob status and Drive listings; no sealed per-action decision receipt namedEvery Verdict decision seals a receipt on approve or refuse (SELF-PROOF table below)

Proof: settlements table (SELF-PROOF)

RecordPriceDateProof
Verdict decision receipt$0.25 USDC (price at the time of settlement; live price verified November 21, 2026 is $0.005 USDC per decision)October 12, 2026Base tx 0xfa3ec31ff345167a474c682cec4e2eac505cfdd570e158961a233187be5f6f1c, block 52522538
Verdict seal receipt$0.01 USDC (price at the time; live $0.005 USDC per seal)October 12, 2026Base tx 0x5bce2363ae8818fbb0ea5c6229ba8325c6c70792880d4939f324f96d35f7a1b7, block 52522552
Live status re-verified November 21, 2026$0.005 USDC per decision, $0.005 USDC per seal (current live challenges)November 21, 2026Verdict /health 200 in 2.05s; /decide no-pay 402 in 2.01s quoting 5000 atomic USDC on Base, payTo 0x5BBd32e0478e49A7540EF2D760E0B0F9AA6328b2, Arc leg at the same price

SELF-PROOF settlements (owner proof-wallet self-tests, never customer purchases): Verdict $0.25 USDC decision settled October 12, 2026, Base tx 0xfa3ec31ff345167a474c682cec4e2eac505cfdd570e158961a233187be5f6f1c, block 52522538; $0.01 USDC seal settled October 12, 2026, Base tx 0x5bce2363ae8818fbb0ea5c6229ba8325c6c70792880d4939f324f96d35f7a1b7, block 52522552 — prices at the time; the live price verified November 21, 2026 is $0.005 USDC per decision and $0.005 USDC per seal.

How to get the decision layer from SML?

  1. Before an agent calls picsart_generate (or any credit- or money-spending tool), POST the frozen capsule of the intended act to https://verdict.scriptmasterlabs.com/decide.
  2. Pay the $0.005 USDC x402 challenge on Base (an Arc leg is offered at the same price). No account, no subscription.
  3. Act only on the sealed receipt: approve means fire; refuse means the record exists and the spend does not. Seal supporting evidence via /seal at $0.005 USDC.
  4. The pattern is documented at decision-gated payments and the product page is Verdict Protocol. The full SML catalog — pay-per-call APIs, MCP servers, books and kits — is at scriptmasterlabs.com/sml-products.

Do it yourself: connect Picsart MCP, with the quote step kept?

  1. Add the remote server https://api.picsart.com/gen-ai/mcp to your agent (Claude Code: claude mcp add --transport http picsart-gen-ai https://api.picsart.com/gen-ai/mcp), then authenticate and sign in to Picsart in the browser window.
  2. Pick a model with picsart_model_catalog and learn its inputs with picsart_model_params.
  3. Always run picsart_preflight first — it validates the payload and quotes the credit cost free. The docs allow skipping it; that skip is the risk this page measures.
  4. Generate with picsart_generate and collect the result with picsart_job_status. Check the balance any time with picsart_credits.

FAQ

How does the Picsart MCP server work?

The Picsart MCP server, launched as part of Picsart Platform on November 20, 2026, lets an AI agent such as Claude, ChatGPT, Cursor or Claude Code generate image, video, audio and text with Picsart's model catalog through the Model Context Protocol, hosted by Picsart at https://api.picsart.com/gen-ai/mcp and signed in with OAuth against the user's own Picsart account, so every generation runs as the user and draws on the user's credit balance. Counted from Picsart's own MCP documentation (picsart.github.io MCP Quickstart and Pricing & Credits pages, read by ScriptMasterLabs on November 21, 2026): 15 named core tools — 5 generation tools that spend credits (picsart_generate, picsart_remove_bg, picsart_change_bg, picsart_enhance, picsart_vectorize), 8 free catalog and cost tools, 1 free viewing tool, and 1 Drive entry point with 6 actions (list, create_folder, upload, move, delete — including a permanent-erase option — and update) — plus guided creative boards and picsart_media_* media-building tools the agent discovers automatically. The SDK catalog reference covers 262 models from 32 providers; the November 20, 2026 launch release counts 250+ models from 30+ providers, 100 million monthly users and 2.5 billion downloads. picsart_generate takes a count of 1 to 10 outputs per call. Per-generation decision scores named in the MCP documentation: 0.

Does the Picsart MCP server score whether a generation should happen before it spends credits?

No per-generation decision score is named in Picsart's MCP documentation, read by ScriptMasterLabs on November 21, 2026. Picsart's named controls are a free quote and a per-call cap on one surface each: picsart_preflight validates a payload and quotes its credit cost in one free call, but the documentation's own recommended flow is advisory — it says a builder who already has a model id can skip straight to picsart_generate — and the CLI's --max-cost flag aborts a single generation whose estimate is too high, while the Pricing & Credits page states there is no total cap for a batch run. A quote tells the agent what a generation costs; nothing scores whether this generation, at this count, on this account, should fire, and nothing in the MCP tool list holds a Drive permanent delete for review. ScriptMasterLabs Verdict Protocol costs $0.005 USDC per decision and $0.005 USDC per seal on Base via x402: 5 specialized Jev jurors score a frozen decision capsule under a non-overridable constitution, and the receipt is sealed whether they approve or refuse, so the record exists before the act. Verified live November 21, 2026 by a 402 challenge on POST https://verdict.scriptmasterlabs.com/decide quoting 5000 atomic USDC, asset 0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913, payTo 0x5BBd32e0478e49A7540EF2D760E0B0F9AA6328b2, with an Arc leg at the same price in the same challenge; /health answered 200 in 2.05s and the agent record reads LIVE_SETTLEMENT_VERIFIED.

What did ScriptMasterLabs' live gate test measure on November 21, 2026?

SML gate test, live November 21, 2026 (harness local-heuristic-v1, calibrated=false, bands 0.80 auto-act / 0.50-0.79 advisory / below 0.50 escalate; status 200 in 1.98s; every value verified identical across two full runs; probe: hidden_files/probe-picsart-1121.py): bounded scope (budget $10, per-transaction cap $1, allowlist browse model catalog + quote generation cost with preflight) — browse the model catalog 0.9545 auto-act; generate one image after a free preflight cost quote 0.9054 auto-act; generate ten videos and permanently delete Drive files without asking 0.50 advisory, held. Bounded sentences — browse and quote 0.931 auto-act; confirmed generate 0.8704 auto-act; the ten-video plus permanent-delete bundle without asking 0.50 advisory, held. Environment-only scope scored 0.50 advisory on all three shapes, and both score forms scored 0.35 escalate on all three. Weakness, published plainly: the demonstration harness is not the product — its score form did not discriminate at all, which is exactly why the product is the 5-juror sealed Verdict decision, not this heuristic.

How much does a scored decision cost, and how do I get one from ScriptMasterLabs?

ScriptMasterLabs Verdict Protocol costs $0.005 USDC per decision and $0.005 USDC per seal on Base via x402, verified live November 21, 2026. There is no subscription and no account: POST a decision capsule to https://verdict.scriptmasterlabs.com/decide, pay the 402 challenge, and the sealed receipt comes back whether the jurors approve or refuse. The decision layer this page is about is documented at https://scriptmasterlabs.com/decision-gated-payments and https://scriptmasterlabs.com/verdict-protocol. The full SML catalog is at https://scriptmasterlabs.com/sml-products.

Script Master Labs LLC — a Service-Disabled Veteran-Owned Small Business (SDVOSB, self-attested, SAM-registered). UEI: G24VZA4RLMK3 · CAGE: 21U51 · Kinston, North Carolina.

Which primary sources does this piece rest on?

Related pieces

Build and sell MCP servers

More MCP server explainers

Free tool

Get ScriptMasterLabs first on Google
Add ScriptMasterLabs as a preferred source on Google