Troubleshooting

My GEDCOM will not upload

Check three things in order. The file must be a GEDCOM with a .ged or .gedcom extension - a .zip from the other platform has to be unpacked first. The export on the other platform must have finished; a download that was interrupted produces a file that ends mid-record. And a tree above the platform ceiling is refused at the door rather than imported halfway. When a file is refused, the job says why on its own page; if the reason there is not clear to you, send it to us with the feedback button.

The import says my tree is larger than my plan allows

The file was read and refused before a single person was written, so nothing was truncated. The file is kept. Upgrade and use Continue this import on the same page: the stored file is imported again without uploading it, and the plan is checked afresh. On the free plan a tree above the plan's threshold is imported in full but only the core runs: the tree, the people and the duplicates search. Hypotheses, the Evidence Graph and research reports for that tree run on Pro.

Open:Pricing

My tree has been processing for a long time

A large GEDCOM takes real time to read, parse and index, and the work is done in stages rather than all at once. The jobs page shows which stage the file is at and how much of it has been read, so you can see it moving. A job that has genuinely failed says so on that page - it does not sit at a stage forever pretending to work.

Open:Upload GEDCOM

My tree says it is not indexed yet

Indexing is a separate stage after the file has been parsed, and the checks need the index before they have anything to measure. Until it finishes, the health tiles honestly report that they have not been computed. This resolves on its own; it is a stage, not an error.

A scan shows nothing - is my tree clean?

No. Read the state, not the number. A tile that says not scanned yet or not computed yet is telling you the engine has not run - that is not the same as a result of zero problems, and the product deliberately refuses to show the two as one thing. The fantasy scan starts on its own after every import, and a run of the checks in Tools runs it again together with the file check, the duplicate scan and the health score. Above 241,169 people the fantasy scan does not run at all, and the tile says so. Archive hints are built during the import itself; if the tile still says not computed yet, import the tree again.

It says "Set yourself as the ROOT on this tree first" and I already did

A ROOT belongs to one tree, not to your account. If you have more than one tree, the ROOT you set on one of them does not carry over to the others, and a kit is linked within a particular tree. Open the tree you are working in, confirm the person who is you is marked as its ROOT, and return to the kit.

I cannot link my DNA kit to a person

Two things have to be true before a kit can be linked: the tree needs a ROOT person, and you need to be in the tree that holds the person you mean. If both are true and the person still cannot be chosen, see the next entry - they may be outside the era in which a consumer test could have been taken.

Open:Kits

The person I want is not in the list

Someone born before 1920, or who died before 1990, could not have taken a consumer DNA test, so they are never offered as the owner of a kit - not as a suggestion and not in the search. This is deliberate. If a living relative is missing instead, check that their year of birth is recorded in the tree: a person with no dates at all is ranked below the plausible candidates rather than being hidden.

My match CSV is not recognised, or the form asks for a GEDmatch kit number

Choose by what the file is, not by which company made it. A match list belongs to one kit: register or select that kit first, then use Import matches on its row, where the kit's company is already suggested. The entry that asks for a GEDmatch kit number takes no file at all; if you are holding a FamilyTreeDNA match list, that is the wrong entry, not a broken one.

The match list import says no matches were recognised

Two refusals look alike and are not. A file whose content is an AutoCluster export is not a match list; upload it on the clusters page instead. A real match list in which no row could be read under the company you chose is refused with the number of rows it saw; pick the company the file actually came from and import it again. In both cases nothing was written to the kit, so there is no empty import to clean up.

My clusters did not compute

Clusters come from an AutoCluster export you upload on the clusters page, or are computed from match lists already attached to a kit. If neither has happened yet, there is nothing to cluster, and the page says so rather than showing an empty result as if it had run.

Open:Clusters

Why does AutoTreeGen not find new DNA matches for me?

Because it never goes looking. Matching happens in the database of the company you tested with; AutoTreeGen analyses the exports you bring from there. When you want the newer matches, export the list again from that platform and import the new file - the counts in an export are a snapshot of that company's database on the day the file was made, and two snapshots do not add up.

Why were my two trees not merged automatically?

Because an automatic merge is how a wrong join becomes permanent. Identity is a research decision, so the product finds the candidates, shows the evidence on each side and lets you decide. Reconcile trees is where that comparison lives, and it previews the outcome before anything is applied.

Open:Compare trees

Why does my tree show conflicting dates?

Because two sources you have disagree, and the product shows you that instead of choosing one quietly. Open the conflict to see which source says what. If one of them is clearly the better record, record that decision; if you cannot tell yet, leaving it open is the correct answer.

Why was my famous ancestor flagged?

A flag marks the generation at which documentary support for the line stops - it does not say the line is false, and it is not a comment on your family's tradition. Open the flag to see which link is unsupported. If you hold a document for that generation, attaching it is exactly what clears the flag.

Why is my tree health grade lower than I expected?

The grade measures how well documented the tree is, not how correct it is. A tree that is entirely right but carries few sources scores low, and that is the grade doing its job. Adding sources, resolving conflicts and linking DNA is what moves it.

Why is my source coverage not 100 per cent?

Coverage counts the claims in your tree that have a source attached to them, and most GEDCOM exports carry far more claims than citations. Full coverage is a research goal rather than a normal starting state; the number is there to show you where the undocumented parts are.

A photo will not upload

On the free plan each person holds one photo, a tree holds two hundred, and a file may be up to five megabytes; the message names which of the three it hit. A cropped copy of an existing photo does not count as a second photo. Paid plans have no photo cap. The file must be an image; a document goes through Reconstruct a tree instead.

Open:Photo evidence

How do I report a problem?

The feedback button in the top bar opens a short form: say whether it is a bug, a suggestion or a feature you wish existed, and describe it. You can attach screenshots, and a screenshot is often the fastest way to show us something. Attach only what you want us to see - a screenshot may include names or DNA from your tree, and it goes to the AutoTreeGen team reviewing feedback, nowhere else.