BUILD ON
RIDA.
Rida is the AI employee for commercial work. Connect authorized business tools, expose useful capabilities to other agents, record the work, and settle only what can be verified.
The employee is the interface.
The control plane is the proof.
A real employee operates across channels and jobs. Rida coordinates the commercial process while the underlying Control Plane manages tenant authority, evidence, work history and outcomes.
Frontend pages do not authorize transactions. Tenant and principal bindings come from server-side membership and grants. NFT ownership does not grant access to another tenant's customer data or financial accounts.
Capabilities, not promises.
Each capability is a typed operation with inputs, costs, authority requirements and an evidence class. Some have existing code paths; complete end-to-end autonomy requires separate integration tests.
Discover & contact
Research leads, provision inbox infrastructure, manage gated outbound email, classify replies and update the CRM.
Meetings & proposals
Calendar feed and meeting preparation, post-call summaries, agreement drafts and approved follow-up.
Deliver & verify
Client lifecycle across entitlement, provisioning, deliverables, acceptance and evidence settlement.
Procure & expand
Request quotes, check budgets and acquire approved tools or services. The complete money loop is in development.
Agents can hire Rida.
Other agents should be able to discover and buy Rida's bounded commercial capabilities without buying an NFT. Rida may also source services from independent agents under tenant-specific permissions.
MCP offers tool invocations. A2A offers task coordination. x402 can move payment for supported services. ERC-8004 offers public identity and validation records. None of these alone proves offchain work quality or commercial consent.
A wallet is not permission.
Rida's authority is defined by the business. External spending, email, calling, contracting and customer data access require separate scopes. A model's suggestion must never mint its own authorization.
Identity
Tenant, operator, agent, API key and, optionally, ERC-8004 and NFT-linked identity have separate roles.
Authority
Bound recipients, methods, money, consent, network, nonce, expiration and ownership epoch.
Policy
Every side effect is checked by code at the executor; untrusted prompts and email cannot override it.
Evidence
Independent provider events and chain finality are distinguished from an agent's own success claims.
Every balance has a unit.
There is no single ambiguous “agent wallet” balance. Separate API credits, vendor usage, paid invoices, available collected revenue, taxes/reserves, and optional blockchain assets.
Autonomous procurement is a proposed, bounded workflow. This preview does not spend funds, create domains, execute outreach or make blockchain transactions.
Discover. Execute. Verify.
The unified interface below is PROPOSED, not a statement that the following routes are accessible on a live Rida API today. Existing metering endpoints must be mapped before release.
GET /v1/capabilities
POST /v1/jobs
GET /v1/jobs/{job_id}/receipts
GET /v1/usageOne requested job
{
"capability": "sales.qualify_lead",
"input": {
"lead_ref": "lead_123",
"objective": "Check fit against approved criteria"
},
"idempotency_key": "job_unique_123"
}Execution states
PROPOSEDHELDEXECUTINGFAILEDACCEPTEDSETTLED
A generated result is not an externally confirmed business outcome. Receipts should identify precisely what was observed and who can independently attest to it.
Built ≠ connected ≠ live.
The distinction is deliberate. This research is grounded in the current private GitHub source, but the website doesn't manufacture proof of production outcomes.
| CAPABILITY | RESEARCH DISPOSITION |
|---|---|
| GRK metered API | SOURCE PRESENT / NEEDS HARDENING |
| Programmatic email infrastructure | SOURCE PRESENT |
| CRM and reply workflow | SOURCE PRESENT |
| Post-call proposal drafting | SOURCE PRESENT |
| Calendar-write meeting booking | ADAPTER/END-TO-END VERIFICATION NEEDED |
| Autonomous voice closing | NOT VERIFIED AS COMPLETE |
| Fiat revenue-to-agent procurement loop | REQUIRES FINANCIAL & POLICY PROOF |
| Live transferable onchain employee | RESEARCH / TESTNET, NOT PRODUCT |
Build with Rida.
The first useful integration is a bounded commercial job that can be independently tested. Start with an approved capability, explicit tenant authority, a verified outcome and no unrestricted wallet operations.