01 · Workflow Supervisor
Owns queues, schedules, budgets, heartbeats, conflict resolution, retries and approval state. Finished artifact: an auditable assignment and execution ledger.
Explore SEOForge
The autonomous team
Four agents with clear jobs, bounded tools and visible handoffs.
The SEOForge team is designed like an operating system: the Supervisor controls priorities and capacity, two specialist pipelines investigate and prepare work, and the Site Experience Engineer turns approvals into validated site-native proposals.
Nothing publishes without approvalPaid product accessNo ranking guarantees
01
Workflow Supervisor
02
Content Growth
Content Growth
03
Search Optimization
Search Optimization
04 · Site Experience Engineer
Every production path ends at human approval
Agent responsibilities
Different specialists create different artifacts.
02 · Content Growth Agent
Owns opportunities, gaps, calendars, briefs, articles, editorial checks, media and conversion purpose. Finished artifact: a complete article package.
03 · Search Optimization Agent
Owns rankings, technical health, decay, refreshes, cannibalization, links, competitors and backlinks. Finished artifact: an evidence-backed optimization proposal.
04 · Site Experience Engineer
Owns framework and theme understanding, code or CMS diffs, builds, screenshots and rollback context. Finished artifact: a validated production proposal.
Real handoff
One decaying guide moves through accountable owners.
- 01
Supervisor detects a priority signal and reserves a bounded run
Evidence, owner and next state remain visible.
- 02
Search Optimization verifies a sustained intent and CTR change
Evidence, owner and next state remain visible.
- 03
Content Growth supplies new source and answer-structure context if needed
Evidence, owner and next state remain visible.
- 04
Site Experience prepares the exact MDX, metadata and internal-link diff
Evidence, owner and next state remain visible.
- 05
Human reviewer accepts, edits or rejects the production proposal
Evidence, owner and next state remain visible.
- 06
Validation records deployment health and starts the observation window
Evidence, owner and next state remain visible.
Operational transparency
See what happened without pretending a model’s hidden reasoning is proof.
- Assignment, objective, owner and allowed tools
- Evidence URLs, provider observations and timestamps
- Costs, reserved capacity, retries and terminal state
- Artifact diff, preview, validation and rollback context
- Human decision and post-deploy measurement annotation
Paid capacity, same integrity
Every plan includes the four-agent operating model and human production boundary. Plans vary sites, tracked queries, content, media, concurrency and developer access.
Agents cannot self-approve, retrieve credentials, bypass branch protection or promise search outcomes.
Questions
Know the boundary before connecting production.
Are these general-purpose chatbots?+
No. Each agent has a bounded operational role, structured artifacts, scoped tools and explicit handoffs.
Can one agent publish another agent’s work?+
No. Production approval remains a separate authenticated human action.
What if two agents want to change one page?+
The Supervisor identifies overlapping scope, serializes or merges the work and preserves one reviewable proposal.
Do I receive all four agents on Starter?+
Yes. Plans change operating capacity, not the core safety model.
Do the agents guarantee rankings?+
No. They prepare and measure controlled improvements; search engines determine distribution.
Paid deployment · human authority
Deploy the complete team—not a disconnected set of AI tools.
Choose a paid plan, connect scoped providers and keep explicit authority over every production change.