Documentation

Keist.ai

Turn one sentence into a structured plan—then execute with your team using chat, boards, and timelines. This guide mirrors how modern docs sites organize content: start with the basics, then go deeper by section.

Overview

Keist.ai is built around a single workflow: you describe outcomes in natural language, the assistant proposes tasks and structure, and you refine until the plan matches reality. From there, work stays connected across chat, lists, and visual boards so nothing lives only in a thread.

Who this is for

Product and engineering teams who want fast planning without losing accountability—whether you are scoping a launch, a sprint, or a cross-team initiative.

Chat-first: plans emerge from conversation, not static forms.
Workspace-scoped: context and history stay tied to the right product or client.
Execution-ready: tasks flow into To-dos, Kanban, and Gantt when dates matter.

Quickstart

Create an account and open the app from the home page or sign in.
Pick or create a workspace so your chats and tasks are grouped for the initiative you are working on.
Describe your goal in one message—for example a roadmap, sprint breakdown, or onboarding checklist.
Iterate in chat until scope, owners, and timing feel right, then move work to boards or the timeline as needed.

Try asking

“Draft a two-week plan for our mobile release with milestones for QA, docs, and marketing.” The model can propose tasks you can edit or split in follow-up messages.

Workspaces

Workspaces isolate conversations, to-dos, and boards by context—such as a product line, customer, or internal program. Switching workspaces keeps search results and history relevant and reduces noise when you juggle multiple streams of work.

When collaboration is enabled for your organization, you can invite others to a workspace so everyone sees the same plans and execution state. Permissions depend on your deployment and admin settings.

Plan mode & chat

The chat is where plans are created and revised. Ask for roadmaps, backlogs, RACI-style ownership, or estimates. Follow up with constraints—“only two engineers,” “must ship before Friday,” “exclude legacy APIs”—and the assistant will adjust without throwing away prior context.

Long outputs can be refined section by section. If something is vague, ask for acceptance criteria or subtasks until each item is actionable.

Views & boards

Beyond chat, the product offers structured views so you can scan and manage work visually. Availability of each view can depend on your workspace configuration.

View	Best for
To-dos	Linear lists, quick status checks, and bulk edits.
Kanban	Workflow columns and drag-and-drop between stages.
Gantt	Time-based planning when tasks have start and end dates.

To-dos & Kanban

Move from agreed plan to day-to-day execution using the To-dos list and Kanban board. Drag cards between columns, update status, and keep titles aligned with what you committed to in chat.

For teams, Kanban makes blockers and throughput visible; To-dos work well for personal triage or simple checklists derived from a plan.

Gantt & timeline

When tasks include dates, the Gantt view shows how work overlaps and sequences over time. Drag bars to shift schedules and align stakeholders on ordering at a glance.

Agents & integrations

Connect tools your team already uses—development, docs, messaging, and more—so plans and updates can sync with systems of record. Available connectors and agent capabilities depend on your workspace; check Settings for what is enabled.

REST API

The external HTTP API lets workspace agents authenticate with a credential and read data such as conversations from your workspace. It is intended for server-to-server or automation use—not for embedding secrets in public clients.

Base URL

All routes are under the v1 prefix. Replace the host with your deployment (production, staging, or self-hosted):

https://<your-host>/api/external/v1

Authentication

Send a Bearer token in the Authorization header. The token is issued when you create an agent credential in the workspace; it has the form ap_<publicId>.<secret>.

Authorization: Bearer ap_<publicId>.<secret>

Invalid or inactive credentials receive 401 Unauthorized. Missing scopes return 403 Forbidden.

Scopes

Each credential lists allowed scopes. The conversations endpoint requires conversations:read.

Endpoints

GET /api/external/v1/ping

Verifies the token and returns workspace and agent identifiers. Useful for health checks and confirming credentials after rotation.

JSON response includes ok, workspaceId, agentId, scopes, and conversationsRead (boolean).

curl -sS \
  -H "Authorization: Bearer ap_<publicId>.<secret>" \
  "https://<your-host>/api/external/v1/ping"

GET /api/external/v1/conversations

Lists recent conversations for the workspace associated with the credential. Requires the conversations:read scope.

Query: limit (optional, default 50, max 100).

Response includes a conversations array with id, title, userId, workspaceId, createdAt, updatedAt (ISO 8601).

curl -sS \
  -H "Authorization: Bearer ap_<publicId>.<secret>" \
  "https://<your-host>/api/external/v1/conversations?limit=25"

Credentials and agents

Create or manage agents and API credentials from your workspace settings (where your deployment exposes them). Store secrets in a vault or environment variables—never commit them to source control.

Evolving surface

Additional endpoints and scopes may ship over time. Prefer checking the latest in-app agent settings or your deployment changelog for your tenant.

Billing & usage

Some capabilities are usage-based. When billing is enabled for your account, you can review balance, top-ups, and usage from billing settings. Pricing may vary by region and product tier.

Deployment-specific

Enterprise or self-hosted deployments may use different billing flows. Follow the guidance from your administrator if it differs from the in-app billing UI.

Security & privacy

Workspace data is handled according to the security practices for the environment you use. For compliance or data residency questions, rely on your organization’s policies or contact support for your deployment.

Review the terms presented at signup and any in-product notices for how retention and model training apply to your account.