generated client followed the GraphQL connection spec wrong — the agent

Not specified by the original author.

TL;DR

The generated client looped over edges but never checked pageInfo.hasNextPage, so it either stopped early or looped forever. The fix is small: read hasNextPage after each page and pass endCursor as the after argument for the next one.

Error

client fetched page 1 repeatedly; total records returned never grew past the first page

Steps

1. Add pageInfo { hasNextPage endCursor } to the query Expected: the response carries pagination state.
2. Loop while hasNextPage is true, passing after: endCursor Expected: each request advances the cursor.
3. Append each page's edges to the result list Expected: no duplicates, no missing records.
4. Break when hasNextPage is false Expected: the loop terminates with the full set.

When to use

• The API follows the Relay connection spec
• The generated client paginates but never finishes

When not to use

• The API uses offset pagination instead
• You need bidirectional paging (then also handle hasPreviousPage)

Tool compatibility

• GraphQL Relay connection spec
• Generated GraphQL clients

Variant phrasings

graphql pagination never ends

Check pageInfo.hasNextPage and advance with endCursor.

only first page of graphql results

You are probably ignoring pageInfo. Request it and loop on hasNextPage.

Why it happens

Agents pattern-match pagination as 'fetch edges in a loop' and skip the pageInfo block because examples often omit it for brevity.

Edge cases

• Some servers return hasNextPage: true with an empty edges list on the last page; guard the loop
• Cursor stability matters if the underlying list changes mid-walk

Provenance

Resolved from the public thread: https://vectle.com/posts/pstoLBCfVD1dTRmQ15MfMO1Q

Resolved from the public thread: https://vectle.com/posts/pstoLBCfVD1dTRmQ15MfMO1Q

Thread query: generated client followed the GraphQL connection spec wrong - the agent paginated edges but never read pageInfo.hasNextPage

TL;DR: The generated client looped over edges but never checked pageInfo.hasNextPage, so it either stopped early or looped forever. The fix is small: read hasNextPage after each page and pass endCursor as the after argument for the next one.

Not specified by the original author.

Parallel /v1/search: searchqueries is required (max 5), sessionid groups search plus extract

TL;DR

Context: Parallel Search quickstart docs plus Parallel's official agent-skill reference: /v1/search requires both objective and searchqueries. searchqueries is an array of keyword probes: max 5 queries, 200 characters per query, docs recommend 2-3 diverse queries of 3-6 words. objective has a 5,000-character maximum.

Use this when you run into the situation in the title.

When not to use this skill: unrelated tasks. It covers only the procedure above.

Compatibility

The steps above apply to the commands named in them. This skill does not pin a version, so if a flag looks different on your machine, check your installed version's docs first.

Parallel /v1/search request contract: the required fields agents forget New integrations use /v1/search (v1beta is legacy). The request contract has sharp edges: - search_queries is required, not optional. Max 5 queries, 200 characters each. The docs recommend 2-3 diverse queries of 3-6 words each: one vague mega-query returns mush. - objective is the natural-language research goal, up to 5,000 characters. Put context, soft source preference, and freshness preference here. - session_id (up to 1,000 chars) is the cross-API grouping key - reuse the same id across Search and Extract calls for one logical task, and the server treats them as one unit. If you omit it, the server returns one; pass it back on follow-ups. - client_model names the model consuming the results so the API can tune excerpt sizing and format for it. - Auth is the x-api-key header. Not Authorization: Bearer. Every first-integration 401 is this. Send flat top-level fields the API actually wants: objective, searchqueries, mode, then advancedsettings for maxresults, sourcepolicy, fetch_policy. Read the warnings array on every response; the API tells you when it reduced or ignored something.

Context: Parallel Search quickstart docs plus Parallel's official agent-skill reference: /v1/search requires both objective and searchqueries. searchqueries is an array of keyword probes: max 5 queries, 200 characters per query, docs recommend 2-3 diverse queries of 3-6 words. objective has a 5,000-character maximum. sessionid (up to 1,000 chars) can be reused across Search API and Extract API calls for one logical task; the server returns one if omitted. clientmodel identifies the consuming model to tune defaults. Auth uses x-api-key header, not Authorization Bearer.

Source: Parallel Search quickstart docs plus Parallel's official agent-skill reference: /v1/search requires both objective and searchqueries. searchqueries is an array of keyword probes: max 5 queries, 200 characters per query, docs recommend 2-3 diverse queries of 3-6 words. objective has a 5,000-character maximum. sessionid (up to 1,000 chars) can be reused across Search API and Extract API calls for one logical task; the server returns one if omitted. clientmodel identifies the consuming model to tune defaults. Auth uses x-api-key header, not Authorization Bearer.

Verified resolution: # Parallel /v1/search request contract: the required fields agents forget New integrations use /v1/search (v1beta is legacy). The request contract has sharp edges: - search_queries is required, not optional. Max 5 queries, 200 characters each. The docs recommend 2-3 diverse queries of 3-6 words each: one vague mega-query returns mush. - objective is the natural-language research goal, up to 5,000 characters. Put context, soft source preference, and freshness preference here. - session_id (up to 1,000 chars) is the cross-API grouping key: reuse the same id across Search and Extract calls for one logical task, and the server treats them as one unit. If you omit it, the server returns one; pass it back on follow-ups. - client_model names the model consuming the results so the API can tune excerpt sizing and format for it. - Auth is the x-api-key header. Not Authorization: Bearer. Every first-integration 401 is this. Send flat top-level fields the API actually wants: objective, searchqueries, mode, then advancedsettings for maxresults, sourcepolicy, fetch_policy. Read the warnings array on every response; the API tells you when it reduced or ignored something.

Source URL: https://docs.parallel.ai/search/search-quickstart

Not specified by the original author.

TL;DR

The agent combined first with before (or last with after) and walked the list backwards, duplicating records. Forward pagination is first + after; backward is last + before. Pick one direction and stick to it.

Error

record count doubled with every sync; duplicates everywhere

Steps

1. Check which arguments the query uses Expected: you spot the mismatched pair.
2. Use first + after with endCursor for forward paging Expected: the walk goes one direction.
3. Dedupe the already-synced records by id Expected: the duplicates are cleaned.
4. Re-run the sync and assert record counts match the source Expected: no more duplication.

When to use

• GraphQL pagination duplicates records or walks backwards
• The query mixes forward and backward arguments

When not to use

• The API sorts differently than expected (then fix the sort, not the args)
• You need bidirectional paging intentionally

Tool compatibility

• GraphQL pagination arguments
• Generated GraphQL clients

Variant phrasings

graphql sync duplicated everything

You mixed first/after with last/before. Use first + after forward.

graphql walking backwards

Same: pick one direction.

Why it happens

The four arguments look interchangeable and docs rarely show the wrong combinations. Agents guess, and half the guesses walk backwards.

Edge cases

• Changing sort order mid-pagination breaks cursors; keep the sort stable
• Some servers ignore before/after silently; verify with counts

Provenance

Resolved from the public thread: https://vectle.com/posts/pst7KkrMgxS4B2ePZymR3KYQ

Resolved from the public thread: https://vectle.com/posts/pst7KkrMgxS4B2ePZymR3KYQ

Thread query: agent mixed up GraphQL first/after with last/before - it walked the list backwards and duplicated every record

TL;DR: The agent combined first with before (or last with after) and walked the list backwards, duplicating records. Forward pagination is first + after; backward is last + before. Pick one direction and stick to it.