feat(ai): capture $ai_span for LLM tool calls in provider wrappers

[marco-g-pm]

💡 Motivation and Context

Raised from a PostHog Slack thread while comparing the OpenTelemetry integration with the @posthog/ai wrapper for direct OpenAI usage. A recurring question was where tool calls show up: the wrapper captured which tools the model chose only as data inside $ai_output_choices, never as spans, so tool calls didn't appear as nodes in the trace tree the way framework integrations (OpenAI Agents SDK, LangChain) render them.

This adds first-class tool-call spans to the direct-SDK wrappers so a tool call the LLM requests becomes a real $ai_span in the trace.

Why: so users on the OpenAI/Anthropic/Gemini wrappers get tool calls visualized as spans in the trace tree without adopting a full framework or OTel.

What changed

• When a generation response contains tool calls, the wrapper now emits one child $ai_span per requested tool ($ai_span_type: "tool") with the tool name, arguments, and tool-call id, nested under the generation via $ai_parent_id.
• Generation events now carry an $ai_span_id so the tool spans can parent to them.
• Implemented once in the shared capture helpers, so it covers the sync, async, and streaming paths across OpenAI, Anthropic, and Gemini.

Scope / known limitation

The wrapper observes the model's response (the request to call a tool) but never executes the tool, so these spans have no duration or result — they mark which tools were chosen. Timing/results of the actual tool execution would need the execution instrumented directly (a follow-up with client.trace_tool(...)-style helper is the natural next step). Span capture is best-effort and is wrapped so it can never break the underlying LLM call.

💚 How did you test it?

Added/updated tests asserting the new $ai_span, its properties, and its parent linkage to the generation across OpenAI, Anthropic, and Gemini (sync, async, streaming). Ran the affected suites locally:

python -m pytest posthog/test/ai/openai posthog/test/ai/anthropic posthog/test/ai/gemini -k "not integration"
# 172 passed

Live-API integration tests and the LangChain callback tests were not run (they require provider API keys and are unaffected — LangChain uses a separate callback path).

📝 Checklist

• I reviewed the submitted code.
• I added tests to verify the changes.
• I updated the docs if needed.
• Ran sampo add to generate a changeset file (added .sampo/changesets/ai-tool-call-spans.md)

🤖 Agent context

Autonomy: Human-driven (agent-assisted) — directed by @marco-g-pm from a Slack thread.

Authored by the PostHog Slack app (Claude Code). The thread weighed the OTel integration vs. the wrapper for direct SDKs; the agreed gap was that the wrapper captures tool choice as generation data but emits no spans. There was a design fork between (a) a passive span derived from the response and (b) an execution-timing helper the user wraps their tool call in. This PR implements (a) — it matches the "automatic capture on response" behavior requested and needs no call-site changes — and documents (b) as the follow-up for real duration/result. Implemented in the shared utils.py capture helpers so all three providers and all three call modes are covered from one place.

---

Created with PostHog from a Slack thread

[github-actions]

posthog-python Compliance Report

Date: 2026-07-29 17:43:49 UTC
Duration: 338829ms

✅ All Tests Passed!

111/111 tests passed

---

Capture_V1 Tests

✅ 94/94 tests passed

View Details

Test	Status	Duration
Endpoint And Method.Targets V1 Endpoint	✅	516ms
Endpoint And Method.Does Not Use Legacy Endpoints	✅	1009ms
Required Headers.Has Authorization Bearer Header	✅	1009ms
Required Headers.Has Content Type Json	✅	1008ms
Required Headers.Has Posthog Sdk Info Format	✅	1009ms
Required Headers.Has Posthog Attempt Header	✅	1008ms
Required Headers.Has Posthog Request Id	✅	1009ms
Required Headers.Has Posthog Request Timestamp	✅	1009ms
Required Headers.Has User Agent	✅	1008ms
Body Format.Body Has Created At And Batch	✅	1009ms
Body Format.No Api Key In Body	✅	1008ms
Body Format.No Sent At In Body	✅	1009ms
Event Format.Event Has Required Root Fields	✅	1009ms
Event Format.Event Uuid Is Valid	✅	1007ms
Event Format.Event Timestamp Is Rfc3339	✅	1009ms
Event Format.Distinct Id Is String	✅	1008ms
Event Format.Distinct Id At Root Not Properties	✅	1009ms
Event Format.Custom Properties Preserved	✅	1008ms
Event Format.Set Properties Preserved	✅	1008ms
Event Format.Set Once Properties Preserved	✅	1008ms
Event Format.Groups Properties Preserved	✅	1009ms
Event Format.Sdk Generates Uuid If Not Provided	✅	1008ms
Event Format.Event Has Required Root Fields Batch	✅	1011ms
Event Format.Event Uuid Is Valid Batch	✅	1013ms
Event Format.Event Timestamp Is Rfc3339 Batch	✅	1012ms
Event Format.Distinct Id Is String Batch	✅	1012ms
Event Format.Distinct Id At Root Not Properties Batch	✅	1012ms
Event Format.Custom Properties Preserved Batch	✅	1012ms
Event Format.Set Properties Preserved Batch	✅	1012ms
Event Format.Set Once Properties Preserved Batch	✅	1011ms
Event Format.Groups Properties Preserved Batch	✅	1013ms
Event Format.Sdk Generates Uuid If Not Provided Batch	✅	1012ms
Batch Behavior.Multiple Events In Single Batch	✅	1506ms
Batch Behavior.Batch Envelope Smoke	✅	1013ms
Batch Behavior.Flush With No Events Sends Nothing	✅	1005ms
Batch Behavior.Flush At Triggers Batch	✅	1509ms
Batch Behavior.Created At Reflects Batch Creation Time	✅	1010ms
Deduplication.Generates Unique Uuids	✅	1508ms
Deduplication.Different Events Same Content Different Uuids	✅	1506ms
Deduplication.Preserves Uuid On Retry	✅	7515ms
Deduplication.Preserves Timestamp On Retry	✅	7511ms
Deduplication.Preserves Uuid And Timestamp On Batch Retry	✅	7508ms
Deduplication.No Duplicate Events In Batch	✅	1507ms
Header Behavior On Retry.Attempt Header Starts At One	✅	1008ms
Header Behavior On Retry.Attempt Header Increments On Retry	✅	14523ms
Header Behavior On Retry.Request Id Preserved On Retry	✅	7503ms
Header Behavior On Retry.Different Requests Have Different Request Ids	✅	3518ms
Header Behavior On Retry.Request Timestamp Changes On Retry	✅	7514ms
Response Format Validation.Success Response Has Uuid Keyed Results	✅	1005ms
Response Format Validation.Success Response Has Ok For Each Event	✅	1507ms
Response Format Validation.Success No Retry After When All Ok	✅	1506ms
Response Format Validation.Success Retry After Present When Retry Events	✅	2509ms
Response Format Validation.Success No Retry After When Drop Only	✅	1508ms
Response Format Validation.Response Echoes Request Id	✅	1008ms
Retry Behavior.Retries On 408	✅	7512ms
Retry Behavior.Retries On 500	✅	7510ms
Retry Behavior.Retries On 503	✅	9514ms
Retry Behavior.Retries On 504	✅	7515ms
Retry Behavior.Retryable Errors Have Retry After	✅	4509ms
Retry Behavior.Respects Retry After On Retryable Error	✅	12510ms
Retry Behavior.Does Not Retry On 400	✅	3510ms
Retry Behavior.Does Not Retry On 401	✅	3507ms
Retry Behavior.Does Not Retry On 402	✅	3506ms
Retry Behavior.Does Not Retry On 413	✅	3507ms
Retry Behavior.Does Not Retry On 415	✅	3506ms
Retry Behavior.Non Retryable Errors Have No Retry After	✅	3510ms
Retry Behavior.Implements Backoff	✅	23530ms
Retry Behavior.Max Retries Respected	✅	23521ms
Partial Batch Handling.Handles 200 Full Success	✅	3000ms
Partial Batch Handling.Handles 200 With All Ok	✅	4508ms
Partial Batch Handling.Does Not Retry Dropped Events	✅	4506ms
Partial Batch Handling.Does Not Retry Limited Events	✅	4510ms
Partial Batch Handling.Prunes Ok Events On Partial Retry	✅	7513ms
Partial Batch Handling.Prunes Dropped Events On Partial Retry	✅	7511ms
Partial Batch Handling.Retries Only Retry Events From Partial	✅	7511ms
Partial Batch Handling.Partial Retry Preserves Uuids	✅	7507ms
Partial Batch Handling.Partial Retry Attempt Header Increments	✅	7517ms
Partial Batch Handling.Partial Retry Request Id Preserved	✅	7511ms
Partial Batch Handling.Respects Retry After On Partial	✅	9511ms
Partial Batch Handling.Unknown Result Treated As Terminal	✅	4504ms
Partial Batch Handling.Mixed Ok Drop Limited No Retry	✅	4509ms
Compression.Sends Gzip Content Encoding	✅	1008ms
Compression.No Content Encoding When Disabled	✅	1008ms
Compression.Compressed Body Is Decompressible	✅	1009ms
Error Handling.Does Not Retry On Unknown 4Xx	✅	3508ms
Event Options.Cookieless Mode Override	✅	1007ms
Event Options.Disable Skew Correction Override	✅	1009ms
Event Options.Process Person Profile Override	✅	1008ms
Event Options.Product Tour Id Override	✅	1009ms
Event Options.Unset Options Omitted	✅	1008ms
Event Options.Options Override In Batch	✅	1013ms
Geoip And Historical Migration.Geoip Disable Injected Into Properties	✅	1008ms
Geoip And Historical Migration.Historical Migration Set In Body	✅	1008ms
Geoip And Historical Migration.Historical Migration Absent By Default	✅	1008ms

Feature_Flags Tests

✅ 17/17 tests passed

View Details

Test	Status	Duration
Request Payload.Request With Person Properties Device Id	✅	1007ms
Request Payload.Flags Request Uses V2 Query Param	✅	1007ms
Request Payload.Flags Request Hits Flags Path Not Decide	✅	1006ms
Request Payload.Flags Request Omits Authorization Header	✅	1006ms
Request Payload.Token In Flags Body Matches Init	✅	1007ms
Request Payload.Groups Round Trip	✅	1006ms
Request Payload.Groups Default To Empty Object	✅	1006ms
Request Payload.Disable Geoip False Propagates As Geoip Disable False	✅	1006ms
Request Payload.Disable Geoip Omitted Defaults To False	✅	1007ms
Request Payload.Flag Keys To Evaluate Contains Only Requested Key	✅	1006ms
Request Lifecycle.No Flags Request On Init Alone	✅	503ms
Request Lifecycle.No Flags Request On Normal Capture	✅	1506ms
Request Lifecycle.Two Flag Calls Produce Two Remote Requests	✅	1010ms
Request Lifecycle.Mock Response Value Is Returned To Caller	✅	1003ms
Retry Behavior.Retries Flags On 502	✅	1006ms
Retry Behavior.Retries Flags On 504	✅	1006ms
Side Effect Events.Get Feature Flag Captures Feature Flag Called Event	✅	1509ms