Orishare
Loyalty, incentives, promotions and engagement infrastructure behind one REST API.
For AI agents
If you are an AI agent adding rewards to an app, start at
Start here for AI agents: sign up with two HTTP calls and no browser, connect
the hosted MCP server at https://api.orishare.com/mcp (Connect MCP), then follow a
recipe for your stack. Autonomy explains what you may apply
yourself. Every page is plain markdown at its URL plus .md; /llms.txt lists them all and
/llms-full.txt holds them in one file.
Connect your AI app
If you are a person who wants Claude, ChatGPT, Cursor, VS Code or Codex to work in your Orishare
account, add the connector URL https://api.orishare.com/mcp/connect to your AI app and sign in,
or create an AI access key. Setup per client: Connect your AI.
What Orishare is
Orishare is API-first. You integrate once by sending customer events; deterministic, versioned rules award points, move tiers, issue rewards, progress challenges and emit webhooks. Non-engineers edit the same rules in the web app, and AI can draft them, but every runtime decision is made by the rule engine from the published version.
How it fits together
- Environments hold everything: a project starts with one test and one production
environment, each with separate keys and data. Names are free-form and you can add more
(Environments and limits). Secret keys (
osk_…) are bound to one environment; personal access tokens (opat_…) manage organizations and name an environment withX-Orishare-Environment. - Events are the only input.
POST /v1/eventsstores the event and returns202, or evaluates inline withmode: syncand returns the effects and a trace. - Rules are JSON documents (
orishare.rule/v1) with a trigger, conditions, limits and actions. Publishing creates an immutable version; the next event uses it. - Programs sit on top: currencies and the ledger, tier programs, rewards, challenges, campaigns, segments and historical counters.
- Webhooks deliver signed events for everything that changed.
Machine-readable entry points
| What | Where |
|---|---|
| OpenAPI 3.1 | /openapi.json here, or https://api.orishare.com/v1/openapi.json |
| Error codes | /errors (every doc_url in an error body points here) |
| Page index for agents | /llms.txt and /index.json |
| All pages, one file | /llms-full.txt; one page as markdown: its URL plus .md |
| MCP (hosted) | https://api.orishare.com/mcp; sign-in connector https://api.orishare.com/mcp/connect |
| CLI | orishare --help (in the Orishare repository; not on npm) |
Start with the quickstart or, if you are an agent, the integration guide.
Autonomy
How each environment applies change sets: auto, limits or confirm. What limits refuses (reward cost caps and nothing destructive), advisories, how an agent proposes and applies, how refusals and undo work, and who changes the policy.
Quickstart
From signup to the first evaluated event with the CLI only, in a few minutes.