Your agent needs a next step and the reason for it.

TaskSaber separates deterministic ranking from the model you choose. Over MCP an authorized agent asks for ranked work, reads the reasons behind every score, and holds a task lease while it acts.

Agents come with Individual and Business, including trials. Create a credential in the app under Settings → Agents & usage, or on Business from the Organization page for the shared workspace, then point your MCP client at the endpoint.

Any MCP clientJSON
{
  "mcpServers": {
    "tasksaber": {
      "url": "https://api.tasksaber.com/mcp",
      "headers": {
        "Authorization": "Bearer <your credential>"
      }
    }
  }
}

The credential comes from Settings → Agents & usage in the app.

Ask, claim, report, complete.

This is one agent loop as the server sees it, with the units each call costs, beside what you see in the app while it happens.

  1. Waiting for the first call…
What you see in the app
92Fix the double-charge on 3 customer accounts
In Focus

The agent has not asked yet. The task sits where a person would see it first.

Calls
0
Units
0
Left this month
5,000

A full loop is 4 units. Individual includes 5,000 units a month and Business 25,000 per person; Free has no agent access. When the month’s units run out, agents pause until the 1st (UTC). Extra use is never charged.

Why put an agent behind a ranking.

An agent handed the whole backlog has to guess. This contract hands it one ranked task, the reasons, and a lease. Each of the three properties below is something you can press.

The same answer, every time

Ranking is arithmetic over the task, the clock and your configuration. Ask twice and the agent gets the same task and the same reasons, so its behaviour is explainable after the fact.

  • next_tasktsk_9f2a · 92
1 answer

Two agents cannot take one task

A claim is a lease with a fencing token. While one agent holds it, a second claim is refused and told when to retry. Writes without the current fence are rejected on the server.

  • claim_taskgranted · fence 7
1 lease held

Asking is cheaper than browsing

Browsing the backlog costs 2 units a page and hands the model everything. Asking costs 1 unit and hands it one task. The server’s own instructions to the client say so.

Browse
2 units · 1 page of up to 100
Ask
1 unit · 1 task
list_tasks ×1

Connect a client.

  1. Create a credentialIn the app on Individual or Business: Settings → Agents & usage, for an agent that works in your personal workspace. On Business, owners and admins can also connect one to the shared workspace from the Organization page. Revoke either in a click.
  2. Point your client at the endpointAdd the JSON above to your MCP client with the credential as the bearer token.
  3. Ask for the next taskThe agent gets the same ranked answer you see, with the reasons, and can claim a lease before it works.

15 tools, and what each one costs.

Agents come with Individual (5,000 units a month) and Business (25,000 a month per person, plus 25,000 per seat shared by the organization's agents); Free has no agent access. When a month's units are used up, agent calls stop until the 1st (UTC). Extra use is never charged. Units are not model tokens; what you save depends on your workload and provider.

MCP tools and units per accepted call
ToolPurposeUnits
next_taskRequest the next ranked task and its reasons.1
explain_taskInspect the factors contributing to a task’s score.1
list_signalRequest the ranked signal list.2
claim_taskRequest a task lease with a fencing token.1
release_taskRelease a held task lease.0
report_progressReport activity while holding a lease.1
complete_taskRequest completion under the current lease.1
report_estimateReport actual time against an estimate.0
find_blockersInspect dependency blockers.1
create_taskCreate a new task, subject to the tenant’s active-task quota.1
update_taskPatch a task’s fields with an optimistic version check.1
add_dependencyDeclare a blocked-by relation; refused if it would form a cycle.1
remove_dependencyRemove a blocked-by relation; a no-op if it did not exist.0
list_tasksBrowse tasks with filters and bounded cursor pagination.2
get_taskRead one task’s full record and current version.1

Access is something you grant.

Scoped permissions

A personal credential works in your personal workspace. An organization credential works in the shared workspace as the owner or admin who connected it, can be limited to teams, never assigns people or deletes tasks, and stops when that person leaves or loses the role. Every tool call is checked on the server against the permissions you gave the credential, and task leases are enforced there too.

Task content can reach your agent

User-selected third-party agents can receive task content. Structured-field redaction does not guarantee privacy for names, financial information or other sensitive material entered in titles and free text.

Check client compatibility

TaskSaber speaks MCP over Streamable HTTP with a bearer credential (endpoint https://api.tasksaber.com/mcp). A client that only runs local servers connects through the mcp-remote bridge. Any MCP client that supports remote servers connects with the configuration above. We publish a client's own instructions after a released build of it has run the ask, claim, report, complete loop against this endpoint; Cursor and Claude Desktop are the first two on that list. Tell us if a client does not connect.

Want an agent to file tasks from a conversation? Give it MCP access above and ask it in your own words; that is the recommended path, and every task it creates passes the same server checks as one you type. There is no bot in Slack listening for work. The Slack shortcut is for people: you pick a message, TaskSaber makes it a task.

The contract, machine-readable.

Everything on this page is generated from one document, served at /.well-known/tasksaber.json. An agent or a build can read it directly; the hash changes when the catalog does.

What it contains

  • Release status and which surfaces are verified live: sales, the agent endpoint, each provider, iOS and Android.
  • The plan catalog: limits, features and trial terms, with hash 450a9945a9be….
  • The 15 tools with units per call, and the monthly allowances.
  • The interface declaration: protocol mcp, and whether the deployment is verified.
GET https://tasksaber.com/.well-known/tasksaber.json

{
  "name": "TaskSaber",
  "status": "owner-configured-release",
  "availability": {
    "salesEnabled": true,
    "agentEndpointVerified": true,
    "verifiedProviders": ["github","jira","slack"],
    "verifiedClients": []
  },
  "catalogHash": "450a9945a9be869de5fb669a6de0eacd2b8e0b621e96bf539b82410d10666555",
  "interface": { "protocol": "mcp" },
  "pricing": {
    "model": "included-units-per-plan",
    "includedUnitsPerMonth": { "individual": 5000, "business": 25000 },
    "organizationAgentUnitsPerSeat": 25000,
    "overageBilling": false
  },
  "catalog": { "agents": { "tools": [ …15 tools… ] }, … }
}
Agents · TaskSaber