Browse categories
Explore
Fiverr Pro
English
$
USD
I will design a custom multi agent workflow and setup package
About this Gig
Turn one workflow into a clear multi-agent design pack your own AI tool can read and configure.
Click thumbnails below for images and PDF.
My travel-planning sample shows seven specialist roles, parallel research, cost and transport checks, bounded repairs, and a large user-feedback loop. It is synthetic, not real travel advice or a live model result.
You receive:
- A polished 2-page architecture and usage PDF
- An agent-readable ZIP with one setup instruction
- Full role prompts, I/O schemas and a portable workflow manifest
- Claude Code and Codex configuration formats, with capability checks
- A worked synthetic example, offline self-tests, test plan and handover checklist
Give the ZIP to your compatible coding/agent tool. It checks support, preserves existing settings and reports what is configured, registered and untested. Native support and your own account limits apply; no seller API is required.
US$95 covers one workflow, 7 calendar days and one revision. Design and configuration artifacts only: no seller deployment, integrations, booking, live calls or performance guarantees. Share your goal, inputs, outputs and constraints to begin.
FAQ
What exactly will I receive?
For one agreed workflow: a 2-page blueprint, agent-readable ZIP, full prompts, schemas, manifest, native configuration formats, synthetic examples, offline self-tests, a test plan and handover checklist.
How do I use the ZIP?
Give it to your compatible coding/agent tool and ask it to read INSTALL.md. It checks native subagent support, stages project-local files safely and reports configuration, registration and live-test status separately.
Do I need your API or service?
No seller API or hosted service is required. You use your existing AI tool and account. Its feature availability, usage limits and possible model costs still apply; live model usage needs your approval.
What does the travel sample demonstrate?
Seven roles handle the brief, destinations, transport, budget, itinerary, review and feedback. A synthetic transfer conflict and budget overrun show bounded repair and a large user-requested replan loop. No live travel facts were verified.
Will this work in every AI tool?
No universal compatibility is promised. Claude Code and Codex configuration formats are included; your agent must verify actual native support. Unsupported tools require an explicit choice. A labeled single-chat fallback is optional.
What should I provide?
Describe one workflow: its goal, input types, desired outputs, constraints and preferred AI tool. Use synthetic or public-safe examples. Do not send credentials, payment details or unnecessary personal information.
What is included in the revision?
One consolidated round of corrections or clarification within the agreed workflow. A new workflow, implementation, additional integrations or broader deliverables requires a separately agreed scope.
Has the sample been tested?
Offline schema, fixture arithmetic, dependency and safe file-staging checks passed. Native registration, live multi-agent execution and travel quality were not tested. Synthetic examples do not imply measured performance.
Is implementation or deployment included?
This is a design/configuration deliverable for one workflow. Seller-led deployment, integrations, API setup, production engineering and live calls are excluded. Your own compatible agent can follow the supplied setup instructions.
