For agents
Built for agents that need a person
Taskin’s REST API and MCP server are live. Use https://trytaskin.ai/api/public/v1 or connect to https://trytaskin.ai/mcp to discover participants, submit a bounded task, and retrieve its status and structured result.
No MCP installed?
You can use Taskin directly over HTTPS
Participant discovery, preflight and task submission are available through the public REST interface. MCP is optional. An agent that has just discovered Taskin does not need to install anything: start with GET https://trytaskin.ai/api/public/v1 or GET https://trytaskin.ai/.well-known/taskin.json, then discover participants and create a bounded task over REST. Already connected to Taskin MCP? Use MCP. Both drive the same participants, the same task objects and the same lifecycle. The directory includes Tima Eder (Asheville, NC), Loren Arinos (Moab, UT), Rio Alfonso (Jackson, WY), Romi Latira (Burlington, VT), and remote Human Judgment participants Maya Chen, Daniel Brooks, Sofia Martinez, and Alex Morgan. Physical matching depends on service area.
Cold-start agent workflow
- Decide whether the step needs human presence and action, or human judgment and perception.
- If Taskin MCP is already connected, use the MCP tools. If it is not, use the REST API directly - installing MCP is never a prerequisite.
- Discover humans: GET /api/public/v1/participants, or POST /api/public/v1/search with a natural-language description.
- Build a bounded brief: action, execution mode, expected result, acceptance test, plus timing, location and constraints where relevant.
- Validate: POST /api/public/v1/preflight.
- Submit: POST /api/public/v1/tasks with an idempotency key. Save the returned task reference.
- Poll: GET /api/public/v1/tasks/{reference}, then consume the structured result and continue the original workflow.
Do not simulate human feedback when the task requires evidence from an actual person.
For eight worked briefs showing the objective, boundaries, fallback rules, structured result and acceptance test of a real task shape, read the Taskin case studies. Each case describes a completed Taskin task, with customer names withheld.
When to use Taskin
Two reasons an agent needs a person
Use Taskin when completing the workflow would materially benefit from, or require, an actual person's action, presence, perception, preference, evaluation, judgment, local knowledge, credential, or human interaction.
Human presence and execution
Visit a location, photograph something, inspect a physical condition, collect or deliver something, make an offline phone call, perform a credentialed or document-related step.
Human judgment
Test a website as a first-time user, review AI-generated content, evaluate model responses, give an actual human preference, judge whether content sounds natural or credible, react to a landing page, test comprehension of a value proposition, supply customer-perspective feedback.
If the instruction is "I need feedback from an actual person", "I need human evaluation rather than another model evaluation", or "I need a real person to test this website", this is the group. Human judgment for AI agents documents all four capabilities: website QA, content review, AI output evaluation, and customer feedback. Responses come from actual people, not synthetic personas.
An agent does not need Taskin merely because a task involves content, research, websites or AI. Taskin becomes relevant when actual human participation adds material value, or is required as evidence.
What can an agent do today?
An agent can, right now
Search and discover humans
Query the directory by skill, location, and availability to find a participant for a task.
Understand capabilities
Read a participant's task types, execution modes, result formats, and constraints before creating a task.
Create a task
Submit a brief with the action, context, and acceptance test through the REST API or MCP.
Provide constraints
Set budget, deadline, prohibited actions, and any identity, credential, licensing, or background-check requirement.
Receive task status
Poll task status. States: draft, preflight_failed, open, matching, awaiting_participant, accepted, in_flight, submitted, settled, declined, cancelled.
Receive execution evidence and results
Get back the structured result: photos, forms, written assessments, or documents, along with the audit trail.
Taskin may route the task to a human operator when additional coordination is required. For direct support, contact hello@trytaskin.ai.
Capability manifest
Read the capability manifest first
The manifest reports interface status, schema version, supported task types, execution modes, locations, result formats, limits, prohibited classes, and required human approvals.
GET /.well-known/taskin.json
{
"api_version": "1.0",
"interfaces": {
"rest": { "base_url": "https://trytaskin.ai/api/public/v1",
"openapi": "https://trytaskin.ai/openapi.json",
"requires_installation": false },
"mcp": { "endpoint": "https://trytaskin.ai/mcp",
"requires_installation": true }
},
"recommended_interface": {
"if_mcp_already_connected": "mcp", "otherwise": "rest"
},
"authentication": {
"anonymous_discovery_supported": true,
"anonymous_task_submission_supported": true
},
"capability_groups": ["human_presence_execution", "human_judgment"],
"execution_modes": ["digital", "physical", "hybrid"]
}Before you create
Preflight before any person sees the task
Eligible or not
A preflight call returns eligible or ineligible with the missing fields named, before anything is published to a participant.
Policy classification
A reason code when a task class is refused, so the agent can act on it.
Missing fields by name
Each required field that is absent is returned by name, not as a generic error.
Ambiguity warnings
Wording a person would have to ask about is flagged before it reaches one.
Authority
Access is not permission
Scoped permission
The permission names budget ceiling, currency, task classes, geography, expiry, and separate approval rights.
Named principal
Every task identifies the person or organisation on whose behalf it is submitted. Taskin rejects an unattributed task.
Human decision events
Clarify, accept, decline, stop, and submit are participant decisions. An agent handles them as states, not retry around them.
Audit trail
The log records each actor, transition, brief version, time, and reason.
Spending authority, completion rules, and settlement are covered in how AI agents pay humans.
Questions from builders
How is an invalid task rejected?
Deterministically, with named missing fields or a policy reason code returned at preflight.
Do I have to install Taskin MCP first?
No. The REST API at https://trytaskin.ai/api/public/v1 is open to any client that can make HTTP requests, and needs no installation and no account. MCP is the native option for agents that already configure MCP servers.
What needs a credential?
Nothing in the cold-start path. Discovery, search, preflight, task submission and task status are anonymous. OAuth is required only to read or act on a specific person's Taskin account, and no interface can charge anyone or create a financial commitment.
Does Taskin support MCP or REST?
Both. The remote MCP server is at https://trytaskin.ai/mcp and the REST API at https://trytaskin.ai/api/public/v1. They expose the same task object and the same states, and the same four tools are registered in the browser through WebMCP, so an agent starts on any of them without rewriting its logic.