Skip to content

fix(devin): support ATIF v1.7 transcripts #526

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
bmcdonough wants to merge 5 commits into
base: main
Choose a base branch
from
Draft

fix(devin): support ATIF v1.7 transcripts #526

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
8ca0871
fix(devin): support ATIF v1.7 transcript format
Jun 19, 2026
8bc9e0a
comment 1, restore lines 52-58
Jun 20, 2026
9c8b8c6
comments 2 & 3, improve wording
Jun 20, 2026
68bc649
comments 6,7,8 ; restore schema_version, agent, steps to be required
Jun 20, 2026
1644fee
changed tool_definiations to an array
Jun 20, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
155 changes: 72 additions & 83 deletions docs/providers/devin.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@ Cognition Devin CLI local usage tracking.

- **Source:** `src/providers/devin.ts`
- **Loading:** eager (`src/providers/index.ts`)
- **Test:** `tests/providers/devin.test.ts` (336 lines)
- **Test:** `tests/providers/devin.test.ts`

## Where it reads from

Expand All @@ -14,7 +14,7 @@ Devin CLI data lives under:
~/.local/share/devin/cli/
```

The MVP usage source is transcript JSON:
Usage comes from transcript JSON:

```text
~/.local/share/devin/cli/transcripts/*.json
Expand Down Expand Up @@ -59,117 +59,106 @@ appear in CLI/UI results until configured.

## Storage format

Transcript root is a JSON object following the [ATIF-v1.4 trajectory schema][atif],
with Devin-specific additions such as per-step `metadata`. The parser does not
validate `schema_version`; it only requires a parseable object with `steps[]`.

Observed [ATIF trajectory schema][atif] variants used by Devin are **ATIF-v1.4** and **ATIF-v1.7**.
The parser does not validate `schema_version`; it only requires a parseable object with `steps[]` array.
Supports Devin-specific additions such as per-step `metadata`.
Core fields include `session_id`, `agent.model_name`, and `steps[]`.

Each counted step can provide:

- `step_id`
- `metadata.committed_acu_cost`
- `metadata.metrics.input_tokens`
- `metadata.metrics.output_tokens`
- `metadata.metrics.cache_creation_tokens`
- `metadata.metrics.cache_read_tokens`
- `metadata.created_at`
- `metadata.generation_model`
- `metadata.request_id`
- `tool_calls[].function_name`

User-input steps (`metadata.is_user_input === true`) are skipped. Non-user
steps are included only if they have positive ACU usage or positive token usage.
### Field normalization

The provider normalizes equivalent usage fields across versions:

- **Cost inputs (ACU before conversion to USD):**
- `metadata.committed_acu_cost`
- `extra.committed_acu_cost`
- `metadata.committed_credit_cost / 10000`
- `extra.committed_credit_cost / 10000`
- **Input tokens:**
- legacy: `metadata.metrics.input_tokens`
- newer prompt-style: `metrics.prompt_tokens - cache_read - cache_creation`
- **Output tokens:**
- `metrics.output_tokens` or `metrics.completion_tokens`
- **Cache write tokens:**
- `metrics.cache_creation_tokens`
- `metrics.cache_creation_input_tokens`
- `metrics.extra.cache_creation_input_tokens`
- **Cache read tokens:**
- `metrics.cache_read_tokens`
- `metrics.cache_read_input_tokens`
- `metrics.cached_tokens`
- `metrics.extra.cache_read_input_tokens`

### User/agent step detection

- User steps are skipped when either:
- `metadata.is_user_input === true` (older exports)
- `source === "user"` (ATIF step source)
- Non-user steps are included only when they contain positive ACU usage or
positive token usage.

### Session, model, and timestamp fallback

- **sessionId:** `session_id` -> `trajectory_id` -> transcript filename
- **model:** `step.extra.generation_model` -> `step.metadata.generation_model` -> `step.model_name` -> `agent.model_name` -> `sessions.model` -> `devin`
- **timestamp:** `step.metadata.created_at` -> `step.timestamp` -> `sessions.last_activity_at` -> `sessions.created_at`

## Pricing

`metadata.committed_acu_cost` is per step, not cumulative. The provider converts
each step with:
`costUSD` is always provider-supplied and uses configured ACU conversion:

@tvcsantos tvcsantos Jun 19, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Still per step and not cumulative, right?


```text
costUSD = committed_acu_cost * devin.acuUsdRate
```

If a step only has `committed_credit_cost`, CodeBurn converts credits to ACU
using Devin's current export convention:

```text
committed_acu_cost = committed_credit_cost / 10000

@tvcsantos tvcsantos Jun 19, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What is the source of this? Can you provide a reference?

Also IMO this should not be hardcoded, but rather configurable. Because Devin can change it at any time, and most likely on enterprise accounts can even depend on the agreed plan.

```

Token-only steps are still included when they have positive token metrics, but
their `costUSD` is `0` if `committed_acu_cost` is absent.
their `costUSD` is `0` if no committed cost is present.

`src/parser.ts` preserves Devin's provider-supplied `costUSD` instead of
re-pricing it through LiteLLM.

## sessions.db enrichment

The provider currently reads these columns from `sessions`:

| Column | Use |
| ------------------- | ----------------------------------------------------------------------------------------------------------- |
| `id` | join key with transcript `session_id` during parsing; discovery uses the transcript filename before `.json` |
| `working_directory` | `projectPath` and derived project name |
| `model` | model fallback |
| `title` | project name fallback |
| `created_at` | timestamp fallback |
| `last_activity_at` | preferred session timestamp fallback |
| `hidden` | skip hidden sessions |

`message_nodes`, `prompt_history`, and `tool_call_state` are not parsed yet.

## Timestamps
The provider reads these columns from `sessions`:

Step timestamps come from `metadata.created_at`, falling back to
`sessions.last_activity_at`, then `sessions.created_at`.

Transcript step timestamps are passed through as ATIF string timestamps.
Numeric normalization is only applied to `sessions.db` timestamps:

- less than `10_000_000_000`: seconds
- otherwise: milliseconds

## Model Resolution

Model names resolve in this order:

1. `step.metadata.generation_model`
2. `step.model_name`
3. `transcript.agent.model_name`
4. `sessions.model`
5. `devin`

## Caching

No provider-level cache.

The normal session cache stores parsed provider calls, but Devin is always
reparsed by `src/parser.ts` because `sessions.db` can change without the
transcript JSON fingerprint changing.
| Column | Use |
| ------------------- | ------------------------------------------------------------------------------------------------------------ |
| `id` | join key with transcript session id during parsing; discovery uses transcript filename before `.json` |
| `working_directory` | `projectPath` and derived project name |
| `model` | model fallback |
| `title` | project name fallback |
| `created_at` | timestamp fallback |
| `last_activity_at` | preferred session timestamp fallback |
| `hidden` | skip hidden sessions |

## Deduplication

`devin:<sessionId>:<step.step_id>`

The provider name is part of the key via the `devin:` prefix.
When `step_id` is missing, parser falls back to 1-based step index.

@tvcsantos tvcsantos Jun 19, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

step_id accoding to ATIF RFC has never been optional and is always mandatory. So it will never be missing


## Quirks

- The transcript directory has usage; `sessions.db` is enrichment only.
- `committed_acu_cost` is per-generation/per-step ACU usage. Never treat it as cumulative.
- Transcript JSON is the usage source; `sessions.db` only enriches metadata.
- There is no default ACU-to-USD rate. Missing config intentionally hides Devin.
- Hidden sessions from `sessions.db` are skipped in discovery and parsing.
- Tool names come directly from `tool_calls[].function_name`; the provider assumes valid ATIF tool-call records.
- If SQLite is unavailable or `sessions.db` cannot be opened, the provider still parses transcripts without enrichment.
- Tool names are taken from either `tool_calls[].function_name` or
`tool_calls[].function.name`.
- If SQLite is unavailable or `sessions.db` cannot be opened, transcript parsing
still works without enrichment.

## When fixing a bug here

1. First check whether `~/.config/codeburn/config.json` contains a valid
`devin.acuUsdRate`. Without it, no Devin sessions should appear.
2. For usage total bugs, compare against:

```bash
jq '[.steps[] | select(.metadata.committed_acu_cost != null) | .metadata.committed_acu_cost] | add' ~/.local/share/devin/cli/transcripts/<session>.json
```

3. If project/model/timestamp metadata is wrong, inspect `sessions.db`, not the transcript.
4. If a hidden session appears, check the `hidden` column. Discovery can only
hide sessions whose transcript filename matches `sessions.id`; parsing uses
the transcript `session_id` when present.
5. Run `tests/providers/devin.test.ts` after parser changes. It covers ACU conversion, disabled-until-configured behavior, timestamp parsing, deduplication, hidden sessions, and `sessions.db` enrichment.
1. Confirm `~/.config/codeburn/config.json` contains a valid positive
`devin.acuUsdRate`.
2. Validate a transcript step's raw usage fields first (cost + token fields).
3. If project/model/timestamp metadata is wrong, inspect `sessions.db`.
4. Run `tests/providers/devin.test.ts` after parser changes.

[atif]: https://github.com/harbor-framework/harbor/blob/main/rfcs/0001-trajectory-format.md
Loading