WORKGANG / MEDIA COPY

SHORT BIO — 139 CHARACTERS
Workgang is a pixel-art agent workspace. Build a crew, try sample missions, and review the handoff. Local demo; no live AI or transactions.

LONG BIO — 300 CHARACTERS
Workgang brings a pixel-art crew to research, treasury checks, and writing. Build an agent, assemble specialists, test a brief, and inspect the receipts. The current release is an interactive local demo with sample data and browser storage. No live AI, wallet connection, or transactions are enabled.

LAUNCH POST 01 — 230 CHARACTERS
Meet Workgang.

Build an agent. Assemble a crew. Give it a mission.

Explore research, treasury, and writing workflows with visible steps and downloadable receipts.

Interactive local demo. Sample data. No live AI or transactions.

LAUNCH POST 02 — 249 CHARACTERS
Scout gathers the questions. Anchor checks the sample budget. Quill prepares the handoff.

Try the Workgang crew, create your own agent, and inspect a mission receipt.

This is a local interactive demo. No live AI, wallet connection, or funds moved.

ARTICLE — 750 WORDS INCLUDING TITLE AND SECTION HEADINGS

Meet Workgang: your agents, working together.

Workgang starts with a practical question: what would it feel like to give a small crew a clear assignment, follow its work, and keep the decision in your hands? Its answer is an agent workspace with a cast of pixel robots, a mission brief, and a record you can inspect. Each operator has a purpose. Every mission ends with something to review.

The current version is an interactive local demonstration. You can create agents, select a crew, run sample missions, explore a training treasury, and export records. The results come from rules running in your browser. No live AI model, external research service, wallet, or blockchain transaction is connected.

Three jobs. One shared brief.

The starting crew brings together Scout, Anchor, and Quill. Scout represents research and evidence: identify the sources supplied in a request, make missing information visible, and distinguish a statement from something verified. In the demo, Scout inventories links in your brief. It does not open those pages or check their contents.

Anchor represents treasury review. It checks a proposed SOL amount against a sample balance and its protected reserve, while keeping the sample approval policy visible. Quill represents writing and synthesis, bringing the request and local checks into a readable handoff. Together, the roles demonstrate a sequence: gather context, examine constraints, then organize the next step.

Select up to three operators for a mission. A short brief is enough to begin, and example prompts offer a starting point for project research, a treasury review, or an agent launch plan.

Characters with a place in the workflow

Three full-body pixel robots take the lead in the interface. A near-black background, thin technical lines, and compact navigation leave room for their silhouettes and names. Coral, cyan, and ivory distinguish the starting crew without turning every surface into a bright control panel.

Choosing a character changes the selected team. Below the crew, the workspace brings the mission brief, staged execution log, and results into view. Mission, Treasury, and Activity tabs keep each kind of work easy to find. The visual personality gives the roles a memorable face; the surrounding interface explains what you can do with them.

A launchpad begins with a purpose

The agent builder is a place to define a new operator. Give it a name, choose a research, treasury, or writing specialty, add instructions, and select a uniform color. Saving adds the character to your local crew, where it can join a sample mission.

Those instructions are saved as configuration. They do not connect an AI model or create autonomous behavior in this release. The builder lets you explore the identity and intended responsibilities of an agent before any production integration exists. A launch plan is a planning exercise here: the preview does not publish an agent to a hosted runtime or launch a token.

Treasury practice with clear boundaries

The Treasury view uses labeled sample assets, a protected operating reserve, and a sample member approval threshold. You can enter a recipient, an amount, and a purpose to prepare an unsigned local payment draft. Validation checks the address format and whether the amount fits the sample available balance.

The result remains a draft record. It does not establish that an address exists, collect member approvals, reserve funds, request a signature, or submit a payment. This makes it possible to examine the preparation and review steps without connecting an account or moving assets.

Keep the handoff, including its limits

Mission receipts separate the supplied request, local checks, assumptions, and open review notes. If a brief asks for current information, the demo can flag that freshness still needs verification. If it asks for an external action, the record states that no such action was performed.

Activity lets you inspect earlier missions and unsigned drafts. Download an individual record or export the collection as JSON. Agents and records stay in this browser's local storage; they are not synchronized to an account. Clearing that storage removes them, so exports provide a separate copy.

A preview with room to develop

Connected models, external tools, wallet authentication, and token launch flows are possible future integrations. They are not enabled in this version, and no release dates are announced here. The current invitation is concrete: build a character, choose a crew, test a brief, and read what comes back.

Workgang gives that process a face and a visible record. The crew prepares the handoff. You review the evidence, recognize what remains unknown, and decide what happens next.

