# Tellseed > Create surveys, feedback forms and research campaigns with your AI agent. Share a branded page, collect answers, see results and export your data. Start free. Tellseed is a survey and data collection product controlled through an MCP-connected AI agent. Use it for customer feedback, questionnaires, research, classification and human evaluation. Contributors answer in a browser without a Tellseed account. The organiser brings the audience. ## Availability and pricing Hosted Free is available now: verify your email and connect your agent. No card required. Pro and Team are planned paid options; online checkout is not open yet. Hosted Free includes 3 campaigns; 1,000 answers a month; Unlimited contributors and reviewers; MCP, review, exports, share cards. Each submitted response to one item counts as one answer. AI client and model provider charges are separate. Self-hosting requires source access; the source repository is currently private. ## Start and connect - [Create a free workspace](https://app.tellseed.com/start): verify email, then connect an agent. - [MCP setup](https://tellseed.com/mcp): hosted endpoint https://app.tellseed.com/mcp; Streamable HTTP; workspace bearer key. Setup options for Cursor, VS Code and Claude Code. Direct hosted Claude Desktop and ChatGPT OAuth connector flows are not implemented. - [Pricing](https://tellseed.com/pricing): current capacity, indicative paid prices and availability. - [About Tellseed](https://tellseed.com/about): operated by Miguel Sanchez, a sole trader in the United Kingdom. - [Contact](https://tellseed.com/contact): hello@tellseed.com. ## Campaign guides - [How to create a campaign with AI](https://tellseed.com/guides/create-campaign-with-ai): Create a survey or data collection campaign with Tellseed and your AI agent. Connect, draft, approve, publish and collect real answers. Start free. - [Create a customer feedback survey with AI](https://tellseed.com/guides/customer-feedback-survey): A practical customer feedback survey guide with a copyable AI prompt, ten example questions, review steps and a plan for acting on the answers. - [Create a classification campaign with answer buttons](https://tellseed.com/guides/classification-campaign): Collect human labels with Tellseed: define categories, use answer buttons, preview the task, publish a contributor link and review the resulting dataset. - [How to write useful survey questions with AI](https://tellseed.com/guides/survey-question-design): Improve an AI-generated questionnaire with neutral wording, clear answer formats, sensible length and a practical review checklist before publishing. - [Plan a research data collection campaign](https://tellseed.com/guides/research-data-collection): Plan and launch a human data collection campaign with Tellseed: define the question, audience, consent, questionnaire, pilot and export for analysis. ## Product facts - Default brief: ten distinct generated questions, ten shown one at a time in sequence. Explicit customer counts and supplied question lists take precedence. Larger pools and shorter sampled rounds are configurable. - Workflow: draft, preview, human approval, publish, share contributor link, collect answers, inspect results and export. - Free-text answers and classification choices are supported. Classification needs explicit response options; otherwise a question uses free text. - Optional name and email fields are disabled by default. When enabled, contributors can leave them blank. Email stays in the owner answers view and is excluded from reviewer views and training exports. - The agent can read a supplied website and apply its palette and image bytes. Tellseed does not automatically fetch a website from a URL. Verify the actual preview before claiming branding is applied. - Campaign owner, approval, reviewer and answers links are private access links. Only the intended contributor link should be distributed publicly. - The connected MCP server supplies the current tool schemas and workflow instructions. Read confirmed status; do not infer publication from approval alone. ## Policies and full reference - [Terms](https://tellseed.com/terms) - [Privacy](https://tellseed.com/privacy) - [Billing](https://tellseed.com/billing) - [Full plain-text instructions](https://tellseed.com/llms-full.txt): the same setup documentation and campaign guides rendered on the public website. # Tellseed MCP server: setup and campaign workflow Source: https://tellseed.com/mcp Connect your AI agent to Tellseed to create and manage surveys, feedback forms, research and classification campaigns. The hosted server uses Streamable HTTP with a workspace bearer key. ## Connect a hosted workspace Start free, verify your email and open Connect in your workspace. Use the install link for Cursor or VS Code, or the copyable Claude Code command. The workspace supplies your own connection details, so you do not need to deploy a server or install Tellseed locally. The MCP endpoint is https://app.tellseed.com/mcp. Authentication uses an Authorization bearer header containing your workspace owner key. An unauthenticated request returning 401 is expected; the public instructions are on this page. Keep the key in your agent’s private connection settings. Do not put it in a campaign description, public link, chat transcript, screenshot or Git repository. Signing out of the browser does not revoke an installed agent. Use the explicit key replacement control if you need to invalidate the old key. ## Choose a client that supports this connection The workspace provides hosted setup options for Cursor, VS Code and Claude Code. Other MCP clients must support Streamable HTTP and a custom Authorization header. Check your client’s current configuration format instead of copying a different app’s configuration blindly. Direct hosted Claude Desktop and ChatGPT connector flows are not implemented: Tellseed does not currently provide the OAuth discovery, consent and token exchange those flows need. A local stdio setup is a separate option for operators who already have source and database access. MCP support is a client feature, not a model-size requirement. The agent needs to discover and call tools, follow their schemas and present the human approval step. Your AI provider’s usage charges are separate from Tellseed’s plan. ## Use the campaign workflow Start by asking the agent to check the workspace and list existing campaigns. Describe the audience, purpose, language, question format, branding and collection settings. Ask for ten questions shown one at a time if that is the experience you want. The agent prepares a draft and returns preview and approval links. Inspect the preview, request changes and approve the finished version yourself in the browser. Then ask the agent to publish the approved version and return the contributor link. Ask for the private answers view, statistics and export separately. Keep private access links out of public messages. In a new conversation, the agent should list campaigns instead of guessing an identifier or assuming the previous chat’s state. Example prompt: Check my Tellseed workspace and list my campaigns. Then create a new 10-question feedback campaign about our community event, shown one question at a time. Keep contact collection disabled. Show me the preview and approval link before publishing. After I approve, publish it and return the contributor link and private answers link separately. ## Let the server describe its tools The connected MCP server exposes tool descriptions and input schemas. These are the current interface contract. The agent should discover the available tools and read the returned status and errors instead of inventing tool names, links or successful actions. For orientation, campaign_list finds campaigns in a fresh session and workspace_status reports workspace state. Campaign creation, draft updates, generation, approval status, publication and answer retrieval are part of the workflow. Approval remains a human step even when the agent handles the surrounding work. Tellseed supports generated content and content you provide, free-text and classification tasks, branded contributor pages, configurable name and email collection, results and exports. Confirm the requested settings in the actual draft before publication. ## If the agent cannot connect or publish For a 401, check the endpoint and workspace key in the client settings. A browser login cookie is not the MCP key. A replaced key stops working, so update the installed connection after replacement. If the agent cannot see tools, inspect the client’s MCP connection status and tool permissions. Reload the connection after changing its configuration. A model claiming it is connected is not evidence that the client discovered the tools. If publication is refused, ask the agent for the current campaign status and the approval link. APPROVED and PUBLISHED are different states. If a change invalidates approval, review the updated draft and approve again. Do not ask the agent to bypass that gate. If the contributor flow has the wrong length, ask the agent to check the generated item count and the number shown per round. If branding is missing, inspect the preview and supply a usable logo URL or exact colours. Contact Tellseed with the non-secret error message if the problem remains. # How to create a campaign with AI Source: https://tellseed.com/guides/create-campaign-with-ai Updated: 2026-09-09 Describe what you want to learn. Your agent builds the campaign in Tellseed, you approve it, and people answer on a page you can share. You do not need to write code. ## What you need before you start Tellseed is a survey and data collection tool controlled through an AI agent. It handles the campaign, contributor page and collected answers. Your agent helps turn your brief into questions and calls Tellseed through the Model Context Protocol, or MCP. Create a free workspace, verify your email and open Connect in the workspace. Choose the setup for your agent. Hosted setup is available for Cursor, VS Code and Claude Code. Other clients need remote MCP support with a bearer header; a chat window alone is not a connection. The hosted Free plan includes 3 campaigns and 1,000 answers per month. An answer is one submitted response to one item, so ten people answering ten questions use 100 answers. Your AI client may have its own subscription or usage charges. ## Give your agent a specific brief Tell the agent who should answer, the decision you want to make, the subject, the language and the answer format. Include a website if you want the agent to use your brand. Ask to see a draft before publishing. A normal starting point is ten questions, shown one at a time. You can request a shorter round, a longer questionnaire or a larger pool with a smaller sample per person. Make the generated count and the number each person sees explicit when those should differ. Example prompt: Use Tellseed to create a customer feedback campaign for my independent coffee shop. I want to decide what to improve next month. Generate 10 clear questions in English and show all 10 to each person, one at a time. Ask about the drinks, service, atmosphere and one improvement. Let people skip a question. Keep name and email collection off. Prepare the draft and show me the preview and approval link before publishing. ## Review the real contributor experience Open the preview returned by the agent. Check the actual wording, answer controls, question count, colours, logo and introductory text. Asking for a brand is a brief; inspecting the preview is how you confirm it was applied. Check that the campaign asks only for information you need. Optional name and email fields can be enabled separately; both are disabled by default. If you collect contact details, tell people why and how you will use them. Email stays in the owner answers view and is excluded from reviewer views and training exports. Tell the agent what to change. Changes that affect the approved campaign may require a new approval. Approval is a human action on the approval page; an agent cannot replace that decision. ## Approve, publish and share the contributor link When you are happy with the draft, approve it in the browser and ask your agent to publish that approved version. Approval and publication are separate steps. Ask the agent to confirm the published status and give you the contributor link. Share the contributor link with your audience through the channels you normally use. People can answer in a browser without installing an AI tool or opening a Tellseed account. Tellseed hosts the collection page; you bring the audience. Keep owner, approval, reviewer and answers links with the people who should have that access. They are not interchangeable with the public contributor link. Example prompt: Check the current campaign status. If my approved version is ready, publish it and confirm the published version. Give me the contributor link to share and the private answers link separately. ## Collect answers and make the next decision Ask your agent for the answers page, campaign statistics and an export. The answers view lets you inspect the collected data; the agent can also retrieve data through MCP. An empty campaign has no responses to summarise, even if it is approved or published. If you ask for a summary, have the agent separate the observations in the collected answers from its interpretation. A small or self-selected group can reveal useful problems without representing every customer. Before a wider launch, complete one test round yourself. Confirm that Next advances through the intended number of questions and that the answers appear in the owner view. A preview is useful for design; a published test is the check for the full collection flow. Example prompt: Show me the campaign statistics and answers link. Summarise the responses collected so far, state how many answers support each theme, and flag missing data. Give me an export I can analyse separately. # Create a customer feedback survey with AI Source: https://tellseed.com/guides/customer-feedback-survey Updated: 2026-09-09 A useful feedback survey starts with a decision. Find out what makes customers return, what gets in their way and which improvement deserves attention first. ## Choose the decision and the audience Pick one decision, such as improving onboarding, changing opening hours or fixing a checkout problem. A survey that asks about everything gives people more work and makes the results harder to use. Decide who should receive it and when. Recent customers can describe the experience they just had. Long-term customers can explain what keeps them returning. Record which group you invited so you can interpret the answers in context. ## Start with this campaign brief Connect Tellseed in your agent, replace the business and audience details, then use the following prompt. The agent can draft the questions and campaign; you still review and approve the result. Example prompt: Create a Tellseed feedback survey for customers who used our bicycle repair shop in the last month. Our decision is which part of the service to improve first. Generate 10 short, neutral questions and show all 10 one at a time. Cover booking, communication, repair quality, price clarity and the most important improvement. Use plain English and avoid collecting contact details. Show me the contributor preview and approval link. Do not publish before my approval. ## Ten questions you can adapt These are example questions, not a validated research instrument. Replace the details with your own service and remove anything that will not change a decision. - What did you need help with when you chose us? - How did you first hear about our service? - What was the booking experience like? - What, if anything, was unclear before your visit? - How well did we keep you informed? - Did the result meet your expectations? Please explain. - What did you think of the explanation of the price? - Which part of the experience worked best for you? - What is the one thing you would change? - Is there anything else you would like us to know? ## Check wording, branding and contact collection Read each question aloud. Split questions that ask about two things at once, such as speed and friendliness. Avoid wording that assumes a good or bad experience. Leave room for someone whose experience does not match your assumptions. If you provide your website for branding, ask the agent to apply the logo and palette, then inspect the preview. A public website may not expose a usable logo, so supply an accessible image or exact colours if the automatic extraction is incomplete. If you want to follow up, make email optional and explain its purpose. Do not describe the survey as anonymous if you collect names or email addresses. Keep the contact request separate from the questions about the service. ## Turn feedback into a small action list Approve the finished draft, ask the agent to publish and share the contributor link. Run a test yourself before sending it to your customer list. Keep the answers link private. Review the raw answers before accepting an AI summary. Ask for repeated themes, examples and the number of responses behind each theme. Do not let the agent invent a satisfaction score from text answers without defining and checking the method. Choose one or two changes and decide how you will tell whether they helped. If you repeat the survey, keep important question wording stable so the comparison is meaningful. Example prompt: Review the collected feedback. Group recurring issues, give the number of answers behind each group and include short examples without contact details. Separate customer statements from your recommendations. Suggest three actions we can evaluate. # Create a classification campaign with answer buttons Source: https://tellseed.com/guides/classification-campaign Updated: 2026-09-09 When you need people to choose a category, give them a clear set of buttons. Classification campaigns turn examples into structured human judgements you can review and export. ## When classification is a useful format Use classification when the result should be a label from a defined set: a support topic, content category, intent or sentiment. The contributor reads an item and chooses an allowed answer, instead of typing a new category each time. Use a free-text survey when you need an explanation in the person’s own words. A button is convenient only if the available choices can represent the answer. Tellseed supports classification labels as buttons, so contributors do not need to type the category name. ## Define labels that different people can apply Write a short definition for every category and include an example near the boundary between similar categories. Decide how to handle ambiguous items, missing context and examples that fit none of the labels. For a support triage task, Billing, Technical problem, Feature request and Other can be a starting point. Test those categories against actual examples before sending the campaign to a larger group. Remove personal information that is unnecessary for the judgement. ## Ask your agent to build the classification task Connect your workspace to an agent, then describe the item pool and allowed labels. State how many items you want to prepare and how many each person should see. Those are two different settings. Example prompt: Use Tellseed to draft a classification campaign for support message topics. Prepare 10 fictional support messages with no personal data. Use the labels Billing, Technical problem, Feature request and Other as single-choice answer buttons. Show each contributor 10 items one at a time. Write a concise definition for each label and let people skip uncertain items. Show me the preview and approval link before publishing. ## Choose the item pool and the round length For a small pilot, prepare ten examples and show all ten to each contributor. For a larger task, you can ask for a pool of one hundred examples with ten selected per round. Have the agent report both settings so a large pool does not accidentally become a long questionnaire. If you need several independent labels per item, explicitly define that requirement and verify coverage in the collected data. A short round does not by itself guarantee balanced coverage or agreement between reviewers. Check the preview for real buttons, exact category names, item content and progress. Avoid using a generated example pool as if it were evidence about real customers. ## Inspect the labels before using the dataset Approve the draft, publish and share the contributor link with the people doing the labelling. Ask your agent for the private answers view and an export containing item identifiers and chosen labels. Check for missing categories, skipped items and examples that repeatedly confuse people. If several people label the same item, examine disagreements before collapsing their labels into a single answer. For training or evaluation data, document the instructions and the campaign version alongside the export. Human labels need a clear interpretation; collecting them does not automatically make them a reliable benchmark. # How to write useful survey questions with AI Source: https://tellseed.com/guides/survey-question-design Updated: 2026-09-09 AI can draft a questionnaire quickly. A short human review makes the difference between questions people can answer and questions that produce confusing data. ## Give every question a purpose Start with the decision the answers should support. For every question, finish this sentence: “I will use this answer to…” If there is no clear ending, consider removing the question. Tellseed defaults to ten generated items and ten shown in a round. Ten is a starting point, not a target to fill at any cost. Ask for fewer if the decision only needs a few answers, or a longer questionnaire when the extra work has a clear purpose. ## Remove leading wording and double questions “How much did you enjoy our excellent service?” assumes a positive experience. “How would you describe your experience?” leaves the answer open. Give the agent examples of the difference and ask it to explain proposed revisions. “Was the booking process fast and easy?” asks about two things. Someone may find it easy but slow. Separate the questions or choose the dimension that matters to your decision. Avoid asking people to recall details they are unlikely to remember. A question about their most recent visit is often easier to answer than one about every visit in the last year. Example prompt: Review this questionnaire before creating it in Tellseed. Flag leading wording, assumptions, double questions, repeated questions and unnecessary personal data. Propose a shorter, neutral version and explain what decision each question supports. Do not invent responses. ## Match the answer format to the task Free text is useful for explanations and unexpected issues. Classification buttons are useful when the allowed categories are known. A fixed set of labels makes counting straightforward but can hide answers that do not fit. For a classification campaign, define each label and decide whether Other or a skip option is needed. For open questions, use plain wording and avoid asking for long essays when a sentence would do. Name and email fields are collection settings, not a reason to add contact requests to every question. Both are disabled by default and optional when enabled. Explain the purpose when collecting contact information. ## Try the full round with a fresh reader Read the contributor page on a phone. Check the introduction, answer controls, question order, progress and completion message. Ask someone unfamiliar with the project to tell you what they think each question means. In Tellseed, review the draft through the preview and approval links. After human approval, ask the agent to publish. Submit a test round and check the answers view before inviting your full audience. If the pilot reveals an unclear question, revise it and record the change. Avoid combining results from materially different wording without acknowledging the difference. ## Keep interpretation tied to the evidence A campaign can collect real answers without producing a representative sample. Report who was invited, who answered and how many responses you analysed. Treat missing answers and skipped questions as missing data. When asking an agent for analysis, request the observations first, then its interpretation and suggested actions. Keep an export so you can inspect the underlying answers independently. # Plan a research data collection campaign Source: https://tellseed.com/guides/research-data-collection Updated: 2026-09-09 Use Tellseed to collect responses for interviews, exploratory research and evaluation tasks. Start with a clear research question and a plan for how the answers will be used. ## Separate the research question from the questionnaire A research question describes what you want to understand. The questionnaire asks people about experiences they can actually describe. “Why do new users leave?” might become questions about the task they attempted, the point where they stopped and what they expected next. Choose whether you need open descriptions, category labels or a review task. Tellseed manages the collection workflow and your agent helps prepare it. It does not recruit a representative panel or validate the research method for you. ## Define participants and explain the purpose Specify who is eligible, how you will invite them and what participation involves. Explain the intended use of their answers in the introduction, including whether you plan to quote or contact them. Collect only the information your work needs. If names or email addresses are unnecessary, leave those settings disabled. If they are optional, make that clear. Follow the review and consent requirements of your own organisation for research that needs them. ## Turn the plan into a campaign draft Give your connected agent the research goal, audience, language, format and round length. Include the exact introduction or consent wording your organisation requires instead of asking the model to invent an approval. Example prompt: Draft a Tellseed research campaign about how freelancers organise project handovers. The audience is freelancers who completed a client project in the last three months. Prepare 10 neutral, open questions and show all 10 one at a time. Ask about their most recent handover, documents, communication, difficulties and what they would change. Do not collect names or email addresses. State that this is exploratory research. Return the preview and approval link for my review before publishing. ## Pilot the wording and collection flow Check the contributor preview with a small group before the main collection. Ask whether the questions make sense, whether any need context and how long a round takes in practice. Do not promise a completion time that you have not measured. After approval and publication, run a complete test round. Verify the intended question count, skip behaviour and saved answers. Keep pilot responses distinguishable in your analysis if they should not be part of the main dataset. ## Export a dataset with enough context to interpret Ask the agent for the answers view, current statistics and an export. Keep the campaign version, question wording, collection period and recruitment method with the data. Check whether each row represents one answer or one participant before calculating percentages. Inspect raw responses and missing values before asking for themes. If the agent proposes a coding scheme, review it on a sample and document any changes. Keep interpretations separate from direct statements in the data. Report the limits of the sample. An exploratory campaign can suggest patterns and new questions; the tool alone does not establish causality or population-level conclusions.