MCP Tools

MCP tools are actions a connected agent can call after you connect GigUp with an API token. The agent can read docs, inspect account context, and perform the same core workflows users do in the dashboard: AI Job Trackers, tracked jobs, proposal templates, and proposal generation.

How Tools Work

1. You ask the agent for help with a GigUp task.
2. The agent discovers which GigUp tools are available.
3. The agent calls the relevant tool for docs, setup, account context, trackers, jobs, or templates.
4. The agent uses the result to answer you or plan the next step.

Available Tools

Tool	Permission	What it helps with	Example user request
list-capabilities	read	Shows every available MCP tool, resource, prompt, endpoint, and docs surface.	“What can GigUp MCP do?”
get-setup	read	Returns cloud MCP connection settings for streamable HTTP clients.	“How do I connect this agent to GigUp?”
get-example	read	Returns practical examples for setup, docs, trackers, jobs, templates, and proposals.	“Show me an example of creating a tracker with MCP.”
list-developer-tools	read	Shows the REST API and MCP surfaces with docs links.	“Should I use API or MCP for this?”
list-api-endpoints	read	Lists REST API endpoint summaries and permissions.	“Which endpoints can list tracked jobs?”
get-api-endpoint	read	Returns detailed metadata for one REST API endpoint.	“Explain the create tracker endpoint.”
read-docs	read	Finds documentation summaries and links.	“Which docs explain MCP authentication?”
get-account-summary	read	Shows the authenticated user, current team, plan limits, and usage.	“Check my tracker and template limits.”
update-ai-match-preferences	write	Updates the AI match preferences used when evaluating imported jobs.	“Only match verified clients with SaaS work and healthy budgets.”
search-ai-job-trackers	read	Lists or searches team AI Job Trackers by name, URL, prompt, or active status.	“Show my active Laravel trackers.”
get-ai-job-tracker	read	Returns one AI Job Tracker with URL, prompt, threshold, team, and profile assignment.	“Open tracker 42 and summarize its settings.”
create-ai-job-tracker	write	Creates a tracker from an Upwork search URL and starts the normal background import flow.	“Create a tracker for senior Laravel jobs above $50/hr.”
update-ai-job-tracker	write	Updates tracker name, prompt, match threshold, profile assignment, or active status.	“Raise this tracker threshold to 80% and pause it.”
delete-ai-job-tracker	write	Removes a tracker from the current team.	“Delete the old React tracker.”
search-tracked-jobs	read	Searches tracked jobs with filters for tracker, status, type, difficulty, client rating, spend, and location.	“Find high-spend fixed-price jobs from my SaaS tracker.”
get-tracked-job	read	Returns one accessible tracked or imported Upwork job with AI and status context.	“Show job 991 and explain why it matched.”
update-tracked-job-status	write	Updates or clears job status, saved cover letter text, and connects used.	“Mark this job applied and save the proposal text.”
search-ai-proposal-templates	read	Lists or searches saved AI proposal templates.	“Show my default proposal template.”
get-ai-proposal-template	read	Returns one AI proposal template.	“Open template 8.”
create-ai-proposal-template	write	Creates a reusable proposal generation prompt.	“Create a concise proposal template for Laravel API jobs.”
update-ai-proposal-template	write	Updates template name, prompt, or default status.	“Make this my default template.”
delete-ai-proposal-template	write	Deletes a saved proposal template.	“Remove my old long-form template.”
generate-job-proposal	write	Generates a proposal for a tracked or imported job using an optional saved template.	“Generate a proposal for job 991 using my default template.”

Choosing the Right Tool

Most users do not need to choose a tool. Ask the agent in normal language and let it pick the right tool. Read-only requests need a token with read. Requests that create, update, delete, or generate content need write.

Use GigUp MCP to show my active AI Job Trackers and find the best Laravel jobs from them.

Safe Usage

• Ask the agent to cite the GigUp docs or endpoint it used.
• Use read tokens when the agent only needs to inspect data.
• Review the agent's plan before applying code changes.
• Use separate API tokens for separate agents or workspaces.
• Keep token permissions as narrow as possible.