Seller tutorial

From code archive to a verified Tetrees EX listing

Follow the same guarded workflow whether you start in the browser, an AI IDE through MCP, or the seller API. Code and preview images finish in the signed-in seller manager; verification prep and private feedback can also be handled through MCP or API.

Open seller console Open MCP guide Open API reference

Before you start

Prepare one release candidate. Tetrees stores uploaded versions as immutable artifacts, so fixes are submitted as a higher semantic version instead of overwriting the reviewed code.

A seller-eligible account and completed payout onboarding for paid listings.
A ZIP no larger than 200 MB, or a public HTTPS GitHub, GitLab or Bitbucket repository.
A semantic version such as 1.0.0, a short description, use case, stack and license settings.
Clean preview images that show the product without secrets, tokens or customer data.
Enough Tetrees Points for the mandatory Verified Listing run; higher prep tiers are optional.

Five steps to publish

1. Create the listing draft

Start in the seller console and describe what the product does before uploading code. Accurate product type, category, language, framework and database values improve search, compatibility guidance and verification.

Use a specific name and a concise, factual summary.
Set the price, license period and live demo URL when applicable.
Add the primary language and stack used by the archive, not only the language of its documentation.

2. Upload an immutable version and previews

The seller manager offers three protected ingestion paths: direct private ZIP, API multipart ZIP fallback, and public Git snapshot. The same archive checks, ownership gate and private storage policy apply to each route.

Choose a version greater than every existing version; a reviewed artifact cannot be replaced in place.
Use direct ZIP for the normal browser path, or import a public repository snapshot.
Upload preview images in the browser. MCP does not accept binary code or images; its readiness response provides the exact continuation URL.

Tetrees seller manager showing direct ZIP, multipart ZIP and public Git ingestion, version entry and archive upload controls — Upload preview images in the browser. MCP does not accept binary code or images; its readiness response provides the exact continuation URL.

Upload preview images in the browser. MCP does not accept binary code or images; its readiness response provides the exact continuation URL.

3. Use MCP or API for guided preparation

An ordinary seller token can inspect owned listings, read the submission checklist, quote all four preparation tiers, run one tier only after an exact point-spend confirmation, and retrieve the private AVCP result.

MCP sequence: list_seller_products → get_seller_submission_requirements → get_seller_prep_options → run_seller_prep → get_seller_verification_report.
Direct API uses the equivalent owner-scoped readiness, prep-quote, prep and report endpoints.
Listing edits, code/images, submit and publish remain browser-only; readiness tells the agent what is missing and where to continue.

3. Use MCP or API for guided preparation

4. Run the mandatory AVCP verification

Verified Listing is the mandatory baseline. Pro Prep, UX Prep and Max Prep add deeper preparation but do not replace the baseline gate. The quoted Point cost is bound to the current immutable version and archive size.

Confirm the exact tier, version and quoted Points before a run starts.
A pass enables the publish step for that version.
A failure keeps the listing in Needs fix and returns normalized failing stages, safe evidence, suggested fixes and next actions.

Tetrees seller verification and preparation panel showing the mandatory Verified Listing tier and three optional higher tiers

5. Publish, then update with a new version

Review the public preview and publish only the version that passed its gate. Future code changes use a higher semantic version, a fresh artifact, and a new verification result so buyers receive exactly what was reviewed.

Resolve every blocking readiness item before submitting.
Check the product page, images, documentation and price before publishing.
For updates, add a changelog, increment the version and repeat upload, prep and publish.

5. Publish, then update with a new version

Choose the route that fits your workflow

All routes use the same owner checks and version-bound verification contract. They differ only in where the seller completes each step.

Seller manager

Complete workflow: create/edit listing, upload ZIP or Git snapshot, add preview images, run prep, read feedback, submit and publish.

MCP in an AI IDE

Guided owner workflow: inspect requirements, quote and confirm prep, read AVCP feedback, then open the signed-in browser continuation for files, images and publishing.

Direct seller API

Automation-friendly JSON contract for readiness, quotes, confirmed prep and reports. Use the documented browser continuation for intentionally restricted mutations.

If AVCP does not pass

Do not upload the same version again or guess at the error. Read the private report, fix the release candidate locally, and submit a higher version.

1.Match each failing stage to its safe evidence and recommended fix.
2.Repair build, test, dependency, documentation or security issues in the original project.
3.Create a new ZIP or Git snapshot, increment the version, and run the mandatory tier again.

Ready to list your code product?

Open the seller console to create a draft, or review the MCP/API contracts before connecting an integration.

Start a new listing

From code archive to a verified Tetrees EX listing

Before you start

Prepare one release candidate. Tetrees stores uploaded versions as immutable artifacts, so fixes are submitted as a higher semantic version instead of overwriting the reviewed code.

A seller-eligible account and completed payout onboarding for paid listings.

A ZIP no larger than 200 MB, or a public HTTPS GitHub, GitLab or Bitbucket repository.

A semantic version such as 1.0.0, a short description, use case, stack and license settings.

Clean preview images that show the product without secrets, tokens or customer data.

Enough Tetrees Points for the mandatory Verified Listing run; higher prep tiers are optional.

Five steps to publish

1. Create the listing draft

Use a specific name and a concise, factual summary.
Set the price, license period and live demo URL when applicable.
Add the primary language and stack used by the archive, not only the language of its documentation.

2. Upload an immutable version and previews

Choose a version greater than every existing version; a reviewed artifact cannot be replaced in place.
Use direct ZIP for the normal browser path, or import a public repository snapshot.
Upload preview images in the browser. MCP does not accept binary code or images; its readiness response provides the exact continuation URL.

3. Use MCP or API for guided preparation

MCP sequence: list_seller_products → get_seller_submission_requirements → get_seller_prep_options → run_seller_prep → get_seller_verification_report.
Direct API uses the equivalent owner-scoped readiness, prep-quote, prep and report endpoints.
Listing edits, code/images, submit and publish remain browser-only; readiness tells the agent what is missing and where to continue.

4. Run the mandatory AVCP verification

Confirm the exact tier, version and quoted Points before a run starts.
A pass enables the publish step for that version.
A failure keeps the listing in Needs fix and returns normalized failing stages, safe evidence, suggested fixes and next actions.

5. Publish, then update with a new version

Resolve every blocking readiness item before submitting.
Check the product page, images, documentation and price before publishing.
For updates, add a changelog, increment the version and repeat upload, prep and publish.

Choose the route that fits your workflow

All routes use the same owner checks and version-bound verification contract. They differ only in where the seller completes each step.

Seller manager

Complete workflow: create/edit listing, upload ZIP or Git snapshot, add preview images, run prep, read feedback, submit and publish.

MCP in an AI IDE

Guided owner workflow: inspect requirements, quote and confirm prep, read AVCP feedback, then open the signed-in browser continuation for files, images and publishing.

Direct seller API

Automation-friendly JSON contract for readiness, quotes, confirmed prep and reports. Use the documented browser continuation for intentionally restricted mutations.

If AVCP does not pass

Do not upload the same version again or guess at the error. Read the private report, fix the release candidate locally, and submit a higher version.

1.Match each failing stage to its safe evidence and recommended fix.

2.Repair build, test, dependency, documentation or security issues in the original project.

3.Create a new ZIP or Git snapshot, increment the version, and run the mandatory tier again.