Short answer
These sound similar but do very different things:
Test Mode — a safe preview. See what Lancer would do without sending anything.
External proposal — Lancer still picks the jobs; your own system writes the proposal via webhooks and the API.
External decision & proposal — your own system both picks the jobs and writes the proposal.
Advanced — power-user controls to edit the AI's underlying prompts.
Test Mode
A sandbox. With Test Mode on, nothing is submitted to Upwork and no connects are spent. On any job you can preview the suitability rating and the exact cover letter Lancer would generate — the right way to tune your Knowledge Base and cover letters before going live. In External proposal mode Test Mode still previews the suitability rating (your system writes the proposals, so there is no cover letter to preview); in External decision & proposal it is unavailable because nothing of Lancer's runs. Full guide: How to Use Test Mode.
External modes
For users who want to plug in their own automation (n8n, Zapier, Make, or code). In a campaign's Integrations tab, Bidding mode has three options. Lancer always finds the jobs and always submits the proposal on Upwork; the mode decides how much of the middle Lancer hands to you:
Standard — Lancer's suitability AI decides which jobs are worth bidding on and Lancer writes the proposals. The default.
External proposal — Lancer's suitability AI still decides. Only the jobs it rates suitable are sent to your webhook, together with the rating and the reason, and your system writes the proposal and submits it through the Lancer API. Your Knowledge Base is still required, because the suitability AI reads it; a cover letter template is not.
External decision & proposal — every job that passes the campaign's filters is sent to your webhook, unscored. Your system decides whether to bid and writes the proposal. Neither a Knowledge Base nor a cover letter template is required, so your filters are your only volume control.
Good to know for both external modes:
Switching a live campaign from Standard to either external mode pauses it, so nothing streams before your integration is ready. Switching between the two external modes keeps the campaign running if it already meets the new mode's requirements; otherwise it pauses until you finish the setup.
The "new job" Slack/Discord notifications turn off for that campaign; the webhook is the delivery channel.
If your system doesn't respond to a job, that lead is simply skipped. There is no timeout and no fallback bid.
Every lead card in the Feed gets a Trigger button that re-sends the job to your webhook, which makes testing easy. Trigger sends the job as it is, so a job Lancer has not scored yet arrives without a rating.
On the Analytics tab, External proposal campaigns count suitable and unsuitable jobs exactly like Standard ones. In External decision & proposal, delegated jobs count toward Jobs Analyzed but not toward the suitable/unsuitable split, because that verdict is made by your own system and never reaches Lancer. Setup guide and payload reference: docs.lancer.app. See also the API, MCP & Integrations section.
Advanced
The campaign's Advanced tab lets power users override the system prompts behind suitability and proposal generation, and review the logs. Most users never need it — your Knowledge Base and cover letter template already control the output. Only touch it if you know exactly what you want to change. In the external modes only the prompts Lancer actually runs are shown (suitability in External proposal, none in External decision & proposal).