Jelliu API documentation

Sends a visitor’s message into a webchat conversation and returns the AI agent’s reply in the same response. The call blocks while the model writes, typically a few seconds.

Choosing the thread. Send visitorId to start or continue the thread keyed webchat_<visitorId>, or conversationId (from a previous response) to continue that exact conversation. At least one is required. If conversationId does not exist in the workspace, a thread keyed by visitorId (or by the id itself) is used, creating a new conversation with a different id. Always keep the conversationId from the latest response. A closed thread is reopened by the next message.

Agent. agentId picks the agent that answers and binds it to a thread that has none yet. It never replaces an agent already bound. Without any agent, a generic assistant answers without inventing business details. A paused agent, a deleted agent or a visitor on the exclusion list get a fixed notice instead of an AI reply.

Messages longer than 4000 characters are truncated to 4000 before processing. A message made only of whitespace passes validation but fails processing, so trim before sending.

Side effects. Stores the visitor’s message and the reply. Calls the LLM (or the agent’s voice-engine chat agent) and may run the agent’s tools and knowledge-base lookups. Each model-written reply spends one AI message of the plan’s monthly allowance. Fixed notices spend none. Broadcasts a realtime update to the dashboard and writes an audit entry.

Idempotency. Not idempotent. A retry stores the visitor’s message again, generates a new reply and spends another AI message.

Webhook events. audit.log_recorded for subscribers that selected it explicitly. See Webhooks.

Access

Required scope: write.
Rate limit: General API (120–600 requests/min per workspace by plan) and the webchat limiter — 20 requests/min per client IP. See Rate limits.
Plan: Available on every plan. Requires an active plan with AI messages left; otherwise 403 BILLING_ERROR.

Sends a visitor's message into a webchat conversation and returns the AI agent's reply in the same response. The call blocks while the model writes, typically a few seconds. **Choosing the thread.** Send `visitorId` to start or continue the thread keyed `webchat_<visitorId>`, or `conversationId` (from a previous response) to continue that exact conversation. At least one is required. If `conversationId` does not exist in the workspace, a thread keyed by `visitorId` (or by the id itself) is used, **creating a new conversation with a different id**. Always keep the `conversationId` from the latest response. A `closed` thread is reopened by the next message. **Agent.** `agentId` picks the agent that answers and binds it to a thread that has none yet. It never replaces an agent already bound. Without any agent, a generic assistant answers without inventing business details. A paused agent, a deleted agent or a visitor on the exclusion list get a fixed notice instead of an AI reply. Messages longer than 4000 characters are truncated to 4000 before processing. A message made only of whitespace passes validation but fails processing, so trim before sending. **Side effects.** Stores the visitor's message and the reply. Calls the LLM (or the agent's voice-engine chat agent) and may run the agent's tools and knowledge-base lookups. Each model-written reply spends **one AI message** of the plan's monthly allowance. Fixed notices spend none. Broadcasts a realtime update to the dashboard and writes an audit entry. **Idempotency.** Not idempotent. A retry stores the visitor's message again, generates a new reply and spends another AI message. **Webhook events.** `audit.log_recorded` for subscribers that selected it explicitly. See [Webhooks](/webhooks). **Access** - **Required scope:** `write`. - **Rate limit:** General API (120–600 requests/min per workspace by plan) **and** the webchat limiter — 20 requests/min per client IP. See [Rate limits](/rate-limits). - **Plan:** Available on every plan. Requires an active plan with AI messages left; otherwise 403 `BILLING_ERROR`.

Authentication

AuthorizationBearer

Workspace API key: jl_ followed by 64 lowercase hex characters, created by the workspace owner in the dashboard (Settings → API Keys) and sent as Authorization: Bearer jl_.... The plaintext is shown once, at creation; Jelliu stores only a SHA-256 hash. A workspace can hold up to 25 active keys.

Scope	GET / HEAD	POST / PUT / PATCH / DELETE	Admin-only routes
`read`	Yes	No	No
`write`	Yes	Yes	No
`full`	Yes	Yes	Yes

Operations restricted to admins or owners reject keys without the full scope with 403, and say so in their description. No key, whatever its scope, can mint or revoke API keys or rotate a webhook secret — that requires a signed-in owner session. A revoked key stops authenticating within about 10 seconds. See Authentication.

Workspace API key: `jl_` followed by 64 lowercase hex characters, created by the workspace owner in the dashboard (**Settings → API Keys**) and sent as `Authorization: Bearer jl_...`. The plaintext is shown once, at creation; Jelliu stores only a SHA-256 hash. A workspace can hold up to 25 active keys. | Scope | GET / HEAD | POST / PUT / PATCH / DELETE | Admin-only routes | | --- | --- | --- | --- | | `read` | Yes | No | No | | `write` | Yes | Yes | No | | `full` | Yes | Yes | Yes | Operations restricted to admins or owners reject keys without the `full` scope with `403`, and say so in their description. No key, whatever its scope, can mint or revoke API keys or rotate a webhook secret — that requires a signed-in owner session. A revoked key stops authenticating within about 10 seconds. See [Authentication](/authentication).

Request

At least one of conversationId or visitorId is required.

messagestringRequired1-5000 characters

The visitor's text. Accepted up to 5000 characters, but only the first 4000 are processed and stored.

conversationIdstringOptionalformat: "uuid"

Id of the conversation to continue, as returned by a previous call. Takes precedence over visitorId when it exists in the workspace.

visitorIdstringOptionalformat: "^[a-zA-Z0-9_-]+$"1-100 characters

Your stable key for the visitor (letters, digits, _, -). Trimmed. An empty string is treated as absent.

agentIdstringOptionalformat: "uuid"

Agent that answers. Must be a non-deleted agent of the workspace, or the call returns 403. Only binds threads that have no agent yet.

metadatamap from strings to stringsOptional

Accepted for forward compatibility; currently ignored and not stored.

Response

The agent's reply.

dataobject

Errors

400

Bad Request Error

401

Unauthorized Error

403

Forbidden Error

429

Too Many Requests Error

500

Internal Server Error